diff --git a/content/admin/backing-up-and-restoring-your-instance/configuring-the-backup-service.md b/content/admin/backing-up-and-restoring-your-instance/configuring-the-backup-service.md index 1cb8891b2a75..03921521cb22 100644 --- a/content/admin/backing-up-and-restoring-your-instance/configuring-the-backup-service.md +++ b/content/admin/backing-up-and-restoring-your-instance/configuring-the-backup-service.md @@ -72,6 +72,73 @@ If you're using a dedicated block device as your backup target, you need to init * Sets it to mount automatically at `/data/backup` on boot.{% ifversion ghes > 3.19 %} * If in a clustered environment, configures the node in `cluster.conf` with the `backup-server` role.{% endif %} +#### Increasing the data partition size + +1. Resize the existing backup disk volume mounted at `/data/backup` using your virtualization platform's tools. +1. Connect via SSH as the `admin` user to the node where the backup disk is mounted. See [AUTOTITLE](/admin/administering-your-instance/administering-your-instance-from-the-command-line/accessing-the-administrative-shell-ssh). +1. Put the appliance in maintenance mode. For more information, see [AUTOTITLE](/admin/administering-your-instance/configuring-maintenance-mode/enabling-and-scheduling-maintenance-mode). +1. Reboot the appliance to detect the new storage allocation: + + ```shell + sudo reboot + ``` + +1. Run the `ghe-backup-disk-extend` command to expand the `/data/backup` filesystem: + + ```shell + ghe-backup-disk-extend + ``` + +1. Ensure system services are functioning correctly, then release maintenance mode. For more information, see [AUTOTITLE](/admin/administering-your-instance/configuring-maintenance-mode/enabling-and-scheduling-maintenance-mode). + +> [!WARNING] +> If the `ghe-backup-disk-extend` command (or a prior automatic check) reports: `ghe_backup_data contains a file system with errors`, you must repair the filesystem before retrying the resize. Do not rerun `ghe-backup-disk-extend` until the check completes cleanly. For recovery instructions, see below. + + +#### Repairing filesystem errors + +If the filesystem check fails during `ghe-backup-disk-extend`, follow these steps to repair it. + +1. Ensure the appliance is in maintenance mode and no background jobs are running: + + ```shell copy + ghe-maintenance -s + ghe-resque-info + ``` + +1. Stop and activate the user volume, then run a forced filesystem check (auto‑answer yes): + + ```shell copy + sudo systemctl stop ghe-backup-disk + VGNAME=$(sudo lvs --noheadings -o vg_name | grep ghe_backup_ | awk '{ print $1 }') + sudo vgchange -ay "$VGNAME" + sudo vgscan --mknodes + sudo fsck -fy /dev/mapper/${VGNAME}-ghe_backup_data + ``` + +1. Retry the resize: + + ```shell copy + ghe-backup-disk-extend + ``` + +1. Remount and verify new size: + + ```shell copy + sudo systemctl start ghe-backup-disk + df -h /data/backup + ``` + +1. Reboot and verify the new size. + + + +1. After the appliance restarts, reconnect via SSH to the node on which the backup disk is mounted, then verify the new size: + + ```shell copy + df -h /data/backup + ``` + #### Detach a backup disk > [!WARNING] diff --git a/content/admin/data-residency/feature-overview-for-github-enterprise-cloud-with-data-residency.md b/content/admin/data-residency/feature-overview-for-github-enterprise-cloud-with-data-residency.md index 7acbdc93237f..7f9fe712d7e7 100644 --- a/content/admin/data-residency/feature-overview-for-github-enterprise-cloud-with-data-residency.md +++ b/content/admin/data-residency/feature-overview-for-github-enterprise-cloud-with-data-residency.md @@ -27,7 +27,6 @@ The following features are currently unavailable on {% data variables.enterprise | Maven and Gradle support for {% data variables.product.prodname_registry %} | Currently unavailable. | [AUTOTITLE](/packages/working-with-a-github-packages-registry/working-with-the-apache-maven-registry) | | {% data variables.product.prodname_marketplace %} | {% data variables.product.prodname_marketplace %}, as a means of searching for, purchasing, and directly installing apps and actions, is unavailable. Ecosystem apps and actions can still be discovered and installed from their source, but they may require modification to work on {% data variables.enterprise.data_residency_site %}. | [{% data variables.product.prodname_actions %} workflows from {% data variables.product.prodname_marketplace %}](#github-actions-workflows-from-github-marketplace) | | Some features currently in {% data variables.release-phases.public_preview %} or {% data variables.release-phases.private_preview %} | Certain features that are in a preview phase on {% data variables.product.prodname_dotcom_the_website %} may not be available on {% data variables.enterprise.data_residency_site %} until GA. | | -| Streaming audit logs to Amazon S3 with OpenID Connect (OIDC) | Currently unavailable. | [AUTOTITLE](/admin/monitoring-activity-in-your-enterprise/reviewing-audit-logs-for-your-enterprise/streaming-the-audit-log-for-your-enterprise) | ## Permanently unavailable features diff --git a/content/admin/monitoring-activity-in-your-enterprise/reviewing-audit-logs-for-your-enterprise/streaming-the-audit-log-for-your-enterprise.md b/content/admin/monitoring-activity-in-your-enterprise/reviewing-audit-logs-for-your-enterprise/streaming-the-audit-log-for-your-enterprise.md index 52fa9388f2b4..708b14c861d8 100644 --- a/content/admin/monitoring-activity-in-your-enterprise/reviewing-audit-logs-for-your-enterprise/streaming-the-audit-log-for-your-enterprise.md +++ b/content/admin/monitoring-activity-in-your-enterprise/reviewing-audit-logs-for-your-enterprise/streaming-the-audit-log-for-your-enterprise.md @@ -152,8 +152,8 @@ From {% data variables.product.prodname_dotcom %}: #### Setting up streaming to S3 with OpenID Connect -> [!NOTE] -> Audit log streaming to S3 with OpenID Connect for {% data variables.enterprise.data_residency %} is currently unavailable. See [AUTOTITLE](/admin/data-residency/feature-overview-for-github-enterprise-cloud-with-data-residency#currently-unavailable-features). +> [!WARNING] +> Streaming exports audit log data to the AWS bucket and region you choose, which can be outside your enterprise's data residency region. Review the destination against your data residency and compliance requirements before enabling streaming. From AWS: @@ -179,18 +179,27 @@ From AWS: } ``` + For {% data variables.enterprise.data_residency %}, replace the `sub` condition with the following, replacing `SUBDOMAIN` with your enterprise's subdomain: + + ```json + "oidc-configuration.audit-log.githubusercontent.com:sub": "https://SUBDOMAIN.ghe.com" + ``` + + The `SUBDOMAIN` value is case-sensitive. Use the lowercase subdomain exactly as it appears in your enterprise's URL, not its display name. Keep the `aud` condition set to `sts.amazonaws.com`. + + If your enterprise's subdomain changes or the enterprise is deleted, remove the old subject from your AWS role trust policies. Creating a new trust policy does not revoke an old one. + * Make note of the Amazon Resource Name (ARN) of the created role. From {% data variables.product.prodname_dotcom %}: {% data reusables.enterprise.navigate-to-log-streaming-tab %} {% data reusables.audit_log.streaming-choose-s3 %} 1. Under "Authentication", click **OpenID Connect**. -1. Configure the stream settings. +1. Configure the stream settings.{% ifversion ghec %} -{% ifversion ghec %} - - Under "Region", select the bucket's region. For example, `us-east-1`; an option for Auto Discovery is also available.{% endif %} - - Under "Bucket", type the name of the bucket you want to stream to. For example, `auditlog-streaming-test`. - - Under "ARN Role" type the ARN role you noted earlier. For example, `arn:aws::iam::1234567890:role/github-audit-log-streaming-role`. + * Under "Region", select the bucket's region. For example, `us-east-1`. An option for Auto Discovery is also available. If you use Auto Discovery, verify the destination bucket's region and role ARN before enabling streaming.{% endif %} + * Under "Bucket", type the name of the bucket you want to stream to. For example, `auditlog-streaming-test`. + * Under "ARN Role" type the ARN role you noted earlier. For example, `arn:aws::iam::1234567890:role/github-audit-log-streaming-role`. {% data reusables.audit_log.streaming-check-s3-endpoint %} {% data reusables.enterprise.verify-audit-log-streaming-endpoint %} diff --git a/content/copilot/concepts/customizations/allowing-safe-commands.md b/content/copilot/concepts/customizations/allowing-safe-commands.md index 36a42149f590..c3a9cae33822 100644 --- a/content/copilot/concepts/customizations/allowing-safe-commands.md +++ b/content/copilot/concepts/customizations/allowing-safe-commands.md @@ -21,7 +21,7 @@ Answering approval prompts keeps you in control, but it can slow you down, espec {% data variables.copilot.assisted_permissions_caps %} is a middle ground between manual approval and granting broad permissions with an allow-all option. It uses a large language model to review permission requests and automatically approve actions it judges to be safe in the context of your task. In a standard interactive session, requests that still require approval are put to you. > [!NOTE] -> {% data variables.copilot.assisted_permissions_caps %} is available in {% data variables.copilot.copilot_cli %}, {% data variables.product.prodname_vscode %}, and the {% data variables.copilot.github_copilot_app %}. +> {% data variables.copilot.assisted_permissions_caps %} is available in {% data variables.copilot.copilot_cli %}, {% data variables.product.prodname_vscode %}, and the {% data variables.copilot.github_copilot_app %}. For accurate billing of safety reviews, use {% data variables.copilot.copilot_cli %} version 1.0.104 or later, or the {% data variables.copilot.github_copilot_app %} version 1.1.32 or later. In {% data variables.product.prodname_vscode %}, it's generally available in version 1.143 or later. ## Approval options @@ -53,7 +53,13 @@ If a safety review is unavailable or cannot evaluate a request, the request foll > [!NOTE] > In autopilot mode and non-interactive runs, requests that still require your approval are denied instead. When the safety review requires approval, its explanation is passed back to {% data variables.product.prodname_copilot_short %} so it can try a different approach. See [AUTOTITLE](/copilot/concepts/agent-capabilities/agent-powered-workflows-and-operations/autopilot). -{% data variables.product.prodname_copilot_short %} selects the model used for the safety review, which can differ from the model you selected for your conversation. If the selected review model is unavailable, the review can use your session's active model instead. If you use your own model provider, the review uses your session's active model. +{% data variables.product.prodname_copilot_short %} selects the model used for the safety review, which can differ from the model you selected for your conversation. By default, the safety review uses a GPT Luna model. If the selected model isn't available to you, {% data variables.product.prodname_copilot_short %} uses another model that has been evaluated for safety reviews. If none of these models is available, the request follows the session's normal permission handling. If you use your own model provider, the review uses your session's active model. + +Unless you use your own model provider, each safety review is a separate model request that consumes {% data variables.product.prodname_ai_credits_short %}, in addition to the {% data variables.product.prodname_ai_credits_short %} your task uses. If you use your own model provider, each safety review is a separate request to that provider. With your own model provider, reviews don't consume {% data variables.product.prodname_ai_credits_short %}, but your provider may charge you for those requests. Requests that are already resolved by your permissions or rules don't need a safety review. + +Each safety review typically uses about 0.1 {% data variables.product.prodname_ai_credits_short %} or less with {% data variables.copilot.gpt_6_luna %}. Your usage depends on how many requests are reviewed and how much context each review evaluates. + +For model rates, see [AUTOTITLE](/copilot/reference/copilot-billing/models-and-pricing). > [!IMPORTANT] > {% data variables.copilot.assisted_permissions_caps %} reduces how often you're prompted, but it doesn't remove your responsibility for what {% data variables.product.prodname_copilot_short %} does. The model can make mistakes and approve dangerous or unintended actions. This feature is not a substitute for sandboxing or access restrictions. As with any automatic approval option, you should review the actions {% data variables.product.prodname_copilot_short %} takes, and consider running it in a sandbox or other isolated environment. For more information, see [AUTOTITLE](/copilot/concepts/copilot-surfaces/copilot-cli#risk-mitigation). diff --git a/content/pull-requests/reference/forks.md b/content/pull-requests/reference/forks.md index 06e42c86f6e3..f0663c97d433 100644 --- a/content/pull-requests/reference/forks.md +++ b/content/pull-requests/reference/forks.md @@ -46,6 +46,8 @@ Each fork can have its own: Repository, organization, and enterprise policies can limit whether repositories can be forked and where forks can be created. For private{% ifversion ghec or ghes %} and internal{% endif %} repositories, access to forks also depends on repository visibility, organization membership, and administrator settings. +On {% data variables.product.prodname_dotcom_the_website %}, if you see “You cannot fork this repository at this time” when trying to fork a public repository to your personal account, an account restriction can prevent you from creating a fork, even if you can sign in. The error alone doesn't identify the cause. If you believe your account has been restricted, you can request a review through the [{% data variables.product.github %} Appeal and Reinstatement form](https://support.github.com/contact/reinstatement). + {% ifversion fpt or ghec %} If you're a member of an {% data variables.enterprise.prodname_emu_enterprise %}, additional restrictions apply to the repositories you can fork. {% ifversion ghec %}{% data reusables.enterprise-accounts.emu-forks %}{% endif %} See [AUTOTITLE](/enterprise-cloud@latest/admin/concepts/identity-and-access-management/enterprise-managed-users){% ifversion fpt %} in the {% data variables.product.prodname_ghe_cloud %} documentation.{% else %}.{% endif %} diff --git a/data/release-notes/enterprise-server/3-19/0.yml b/data/release-notes/enterprise-server/3-19/0.yml index c4c0619a03c4..f242bc549b9e 100644 --- a/data/release-notes/enterprise-server/3-19/0.yml +++ b/data/release-notes/enterprise-server/3-19/0.yml @@ -213,6 +213,10 @@ sections: - | YJIT, Ruby's Just-In-Time (JIT) compiler, is enabled by default. Users may experience faster application performance and improved resource efficiency across their instance. [Updated: 2025-12-17] + bugs: + - | + {% data reusables.release-notes.2026-11-saml-issues-non-org-members-fixed %} [Updated: 2026-11-17] + known_issues: # INCLUDE NOTES FOR RELEASE FROM "GHES Release Note Tracking" PROJECT'S "Known Issues" TAB - | diff --git a/data/release-notes/enterprise-server/3-19/15.yml b/data/release-notes/enterprise-server/3-19/15.yml index 48158c2f9fa6..e1414d560c5c 100644 --- a/data/release-notes/enterprise-server/3-19/15.yml +++ b/data/release-notes/enterprise-server/3-19/15.yml @@ -1,16 +1,17 @@ date: '2026-11-10' +intro: | + > [!WARNING] + > The Git security fixes previously listed for this release were incomplete. To receive these fixes, upgrade to {% data variables.product.prodname_ghe_server %} [3.19.16](#3.19.16) or a later patch release of 3.19. [Updated: 2026-11-17] sections: security_fixes: - | - **CRITICAL**: The Ruby Active Storage component was updated to address the file-read and remote code execution vulnerability [CVE-2026-66066](https://www.cve.org/CVERecord?id=CVE-2026-66066). + Packages have been updated to the latest security versions. - | - **HIGH**: Updated Git to include fixes that address [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469). + **CRITICAL**: The Ruby Active Storage component was updated to address the file-read and remote code execution vulnerability [CVE-2026-66066](https://www.cve.org/CVERecord?id=CVE-2026-66066). - | **MODERATE**: The Brotli compression library bundled with the GitHub application was updated to address [CVE-2026-111497](https://www.cve.org/CVERecord?id=CVE-2026-111497). - | Security updates have been applied to Go, OpenSSL, and Ubuntu system packages, including SQLite, across targeted main and bundled container images. These updates partially address [CVE-2026-111943](https://www.cve.org/CVERecord?id=CVE-2026-111943), [CVE-2025-15467](https://www.cve.org/CVERecord?id=CVE-2025-15467), and [CVE-2025-6965](https://www.cve.org/CVERecord?id=CVE-2025-6965); remaining components will be updated in a future release. - - | - Packages have been updated to the latest security versions. known_issues: - | When users clone or fetch a repository over SSH with compression enabled, the operation can stall until the connection times out, especially for repositories with many references. To work around this, users can disable SSH compression or use HTTPS. [Updated: 2026-11-13] @@ -46,3 +47,6 @@ sections: The setting to define private registries at the organization level for code scanning is only available if Dependabot is also enabled for the instance. - | After a restart, an instance can remain unavailable with "Starting..." displayed in the web interface. To recover, an administrator can connect using administrative SSH and rerun `ghe-config-apply`. This provides temporary recovery but does not prevent the issue from recurring. This issue is resolved in {% data variables.product.prodname_ghe_server %} 3.21.0 and later. + errata: + - | + The security fix note for [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469) has been removed because the Git security fixes were incomplete in this release. These fixes are included in {% data variables.product.prodname_ghe_server %} 3.19.16. [Updated: 2026-11-17] diff --git a/data/release-notes/enterprise-server/3-19/16.yml b/data/release-notes/enterprise-server/3-19/16.yml new file mode 100644 index 000000000000..9e05c5bbced5 --- /dev/null +++ b/data/release-notes/enterprise-server/3-19/16.yml @@ -0,0 +1,128 @@ +date: '2026-11-17' +sections: + security_fixes: + - | + Packages have been updated to the latest security versions. + - | + **HIGH:** Updated Git to include fixes that address [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469). This update was previously communicated in the release notes for the last patch release, but had not been fully resolved. + - | + **HIGH:** On a GitHub Enterprise Server instance in a cluster configuration, an attacker authenticated as a site administrator could execute arbitrary commands as the host `git` user on peer Git server nodes outside the pre-receive hook sandbox by supplying a crafted download URL for a custom pre-receive hook environment. GitHub has requested CVE ID [CVE-2026-112740](https://www.cve.org/cverecord?id=CVE-2026-112740) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** A site administrator could modify host files with root privileges and disclose private repository data due to improper path validation in pre-receive hook environment setup. To mitigate this issue, GitHub has strengthened path validation. GitHub has requested CVE ID [CVE-2026-116637](https://www.cve.org/cverecord?id=CVE-2026-116637) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An organization member could read and modify another user's private repositories due to improper SSH certificate identity validation. Exploitation required an instance using built-in authentication and a valid certificate issued to the attacker by a trusted SSH certificate authority. Access was limited to repositories within that authority's scope. GitHub has requested CVE ID [CVE-2026-116643](https://www.cve.org/cverecord?id=CVE-2026-116643) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An attacker with a session for a user whose numeric ID matched an enterprise's ID could generate private keys for its GitHub Apps by bypassing ownership checks. Repository access depended on the App's installations and permissions. On standard fresh installations, the collision involved the non-loginable ghost account; the reported reproduction required site administrator impersonation. GitHub has requested CVE ID [CVE-2026-110307](https://www.cve.org/cverecord?id=CVE-2026-110307) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH**: An attacker with permission to manage custom-property definitions and edit values could add an SSH key to another member's personal account through stored HTML in a single-select option. Exploitation required a definition manager in the same organization to attempt to remove the in-use option and save with a valid sudo session. GitHub has requested CVE ID [CVE-2026-131800](https://www.cve.org/cverecord?id=CVE-2026-131800) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/) + - | + **HIGH**: An OS command injection vulnerability in the GitHub Pages deployment process could allow arbitrary command execution as the GitHub Pages service user on a GitHub Enterprise Server instance. Exploitation required an authenticated user with permission to deploy a GitHub Pages site. GitHub has requested CVE ID [CVE-2026-130824](https://www.cve.org/cverecord?id=CVE-2026-130824) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **MEDIUM:** A user with write access could bypass branch protections through specially crafted input when resolving pull request conflicts. GitHub has requested CVE ID [CVE-2026-131647](https://www.cve.org/cverecord?id=CVE-2026-131647) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **LOW:** SAML authentication could match an incorrect account when NameID values differed only in case, accents, or trailing spaces. These mismatches now return a login-conflict error. + bugs: + - | + On instances configured with multiple trusted IP addresses or CIDR ranges for the Prometheus metrics endpoint, `ghe-config-check` reported validation errors for valid comma-separated lists. + - | + On instances with legacy metrics enabled, the weekly cleanup did not remove metrics files that had not been updated for more than 90 days. These stale files continued to consume space on the root disk. + - | + Site administrators could not apply configuration changes when the backup schedule (`backup.schedule`) was set to a valid cron schedule, such as `0 0 * * *`, because configuration validation rejected whitespace and `*`. + - | + Administrators restoring backups on standalone instances could find that Redis loaded existing append-only file (AOF) data instead of the restored backup. Redis restores could also fail for backups in RDB version 11 format. + - | + Site administrators could not apply configuration after upgrading an instance with a custom OpenTelemetry pipeline that referenced the built-in `otlp/ghes` receiver. + - | + Site administrators whose GitHub Actions storage settings included values such as "github" saw those values incorrectly redacted from hostnames and database names in support bundles, resulting in false diagnostic failures. + - | + After OpenTelemetry metrics were disabled, collectd could stop receiving application and service metrics until it was restarted. + - | + Administrators running `ghe-live-migrations` could see duplicate timestamps and stray formatting characters in the output. Sending SIGTERM to the utility did not always stop the underlying migration container cleanly, and the utility did not reject some invalid options. + - | + When an administrator changed `redis.io-threads` and applied the configuration, Redis on the primary could restart with the previous value, so the change did not take effect. + - | + After a site administrator applied a hotpatch, some files that the hotpatch installed, such as command-line utilities, could be missing from the instance. + - | + Upgrades and other operations that drain a node could take about 10 seconds longer than necessary because an internal service did not shut down promptly when signaled. + - | + After an upgrade, `ghe-config-apply` verified the same Docker images more than once, unnecessarily extending the configuration run. + - | + On standalone or high-availability instances, repository data could be restored even when the corresponding repository metadata was missing from the database backup. Newly created repositories could then be preinitialized with that orphaned data. + - | + Administrators configuring repository caches on instances with GitHub Actions enabled saw a `mssql-server and mysql-server should have the same value` warning during replication setup or configuration runs. + - | + On instances with a large number of committers, enabling security features on a repository, including by applying a security configuration, could time out while calculating remaining GitHub Advanced Security licenses. This calculation is now skipped when enabling the repository uses no additional licenses. + - | + If calculating {% data variables.product.prodname_GH_advanced_security %} license usage took too long while an enterprise owner applied a security configuration, the confirmation dialog did not show how many repositories the configuration would apply to. + - | + When site administrators added or re-added a repository cache replica, repositories could remain unavailable on the cache, causing requests routed to it to return `Repository not found` errors until the background scheduler repaired them. + - | + Site administrators could encounter an error when viewing pending {% data variables.product.pat_v2 %} requests associated with a deleted token. + - | + Organizations could receive email notifications about pending {% data variables.product.pat_v2 %} requests after the associated tokens had been deleted. + - | + On instances running in private mode with subdomain isolation enabled, users could intermittently encounter broken avatar images and failures when loading content from the media, notebooks, and viewscreen subdomains. + - | + Secret scanning alerts for enterprise-level custom patterns did not include links to the custom pattern settings. + - | + For teams linked to SCIM external groups, the site administrator dashboard displayed a **Sync** action that called a service unavailable on GitHub Enterprise Server and a description that incorrectly indicated membership would reconcile automatically. Site administrators can use the **Reconcile** action to reconcile team membership. + - | + Background processes handling issue and project updates could consume increasing amounts of memory. + - | + When GraphQL requests timed out, {% data variables.product.prodname_ghe_server %} could include all executed SQL queries in exception diagnostics, causing logs to grow enough to exhaust the root disk. These queries are no longer included in exception diagnostics. + - | + Users could be incorrectly classified as dormant despite recent activity, such as commenting on issues or pull requests. Last activity timestamps in user reports and the dormant users page could also be outdated. + - | + On instances with high push activity, users could experience delays before repository-related updates appeared in the UI. {% data variables.product.prodname_ghe_server %} processes repository push events with increased parallelism to reduce these delays. + changes: + - | + On AWS, the root volume for new instances launched from the GitHub Enterprise Server Amazon Machine Image (AMI) defaults to the gp3 volume type with 250 MiB/s throughput, instead of gp2. + - | + On clustered instances, the `ghe-cluster-balance` utility no longer displays an outdated disclaimer that described it as a preview feature and directed administrators to contact GitHub Support. The utility has been generally available and fully supported since {% data variables.product.prodname_ghe_server %} 3.9. For more information, see [AUTOTITLE](/admin/monitoring-and-managing-your-instance/configuring-clustering/rebalancing-cluster-workloads). + - | + The `ghe-backup-disk-extend` command-line utility is available to increase the attached disk volume size used by the backup service. + - | + Site administrators can run pre-upgrade live migrations in the background with `ghe-live-migrations --all --background` and check their status with `ghe-live-migrations --status`, which shows whether a migration is running and the latest lines of its log. When run from an interactive terminal, `ghe-live-migrations --all` prompts administrators to choose whether to run in the background. + - | + Redis usage diagnostics use more responsive and reliable keyspace scans, including on larger installations. Collection is bounded by a timeout so diagnostics can continue when Redis scans take too long. + - | + The bundled GitHub Actions runner has been updated to version 2.338.2. + - | + On instances where calculating GitHub Advanced Security license usage takes too long, applying a security configuration to repositories can fail. With help from GitHub Support, site administrators can skip license seat checks when security features are enabled. + known_issues: + - | + During an upgrade of GitHub Enterprise Server, custom firewall rules are removed. If you use custom firewall rules, you must reapply them after upgrading. + - | + If the root site administrator is locked out of the Management Console after failed login attempts, the account does not unlock automatically after the defined lockout time. Someone with administrative SSH access to the instance must unlock the account using the administrative shell. For more information, see [AUTOTITLE](/admin/administering-your-instance/administering-your-instance-from-the-web-ui/troubleshooting-access-to-the-management-console#unlocking-the-root-site-administrator-account). + - | + {% data reusables.release-notes.large-adoc-files-issue %} + - | + Admin stats REST API endpoints may time out on appliances with many users or repositories. Retrying the request until data is returned is advised. + - | + When following the steps for [AUTOTITLE](/admin/monitoring-and-managing-your-instance/configuring-clustering/replacing-a-cluster-node#replacing-the-primary-database-node-mysql-or-mysql-and-mssql), running `ghe-cluster-config-apply` might fail with errors. If this occurs, re-running `ghe-cluster-config-apply` is expected to succeed. + - | + Running a config apply as part of the steps for [Replacing a node in an emergency](/admin/monitoring-and-managing-your-instance/configuring-clustering/replacing-a-cluster-node#replacing-a-node-in-an-emergency) may fail with errors if the node being replaced is still reachable. If this occurs, shut down the node and repeat the steps. + - | + {% data reusables.release-notes.2024-06-possible-frontend-5-minute-outage-during-hotpatch-upgrade %} + - | + When restoring data originally backed up from a 3.13 or greater appliance version, the Elasticsearch indices need to be reindexed before some of the data will show up. This happens via a scheduled job that runs on the first Saturday of each month. It can also be forced by running `/usr/local/share/enterprise/ghe-es-search-repair`. + - | + When enabling automatic update checks for the first time in the Management Console, the status is not dynamically reflected until the "Updates" page is reloaded. + - | + When restoring from a backup snapshot, a large number of `mapper_parsing_exception` errors may be displayed. + - | + When initializing a new GHES cluster, nodes with the `consul-server` role should be added to the cluster before adding additional nodes. Adding all nodes simultaneously creates a race condition between nomad server registration and nomad client registration. + - | + In a cluster, the host running restore requires access to the storage nodes via their private IPs. + - | + After a restore, existing outside collaborators cannot be added to repositories in a new organization. This issue can be resolved by running `/usr/local/share/enterprise/ghe-es-search-repair` on the appliance. + - | + After a geo-replica is promoted to be a primary by running `ghe-repl-promote`, the actions workflow of a repository does not have any suggested workflows. + - | + When publishing npm packages in a workflow shortly after restoring from a backup to GitHub Enterprise Server, you may encounter a `401 Unauthorized` error from the GitHub Packages service. This has been confirmed on earlier releases and may also affect later supported versions. Applicability to current releases is still being confirmed. This can happen if the restore is from an N-1 or N-2 version and the workflow targets the npm endpoint on the backup instance. To avoid this issue, ensure the access token used for publishing is valid and includes the correct scopes for the GitHub Packages service. + - | + The setting to define private registries at the organization level for code scanning is only available if Dependabot is also enabled for the instance. + - | + After a restart, an instance can remain unavailable with "Starting..." displayed in the web interface. To recover, an administrator can connect using administrative SSH and rerun `ghe-config-apply`. This provides temporary recovery but does not prevent the issue from recurring. This issue is resolved in {% data variables.product.prodname_ghe_server %} 3.21.0 and later. + - | + `ghe-storage-extend` will fail for disks with a sector size greater than 1KiB, most notably with Azure Premium SSD v2 disks. This is due to an upstream bug with `resize2fs` when resizing a mounted disk. diff --git a/data/release-notes/enterprise-server/3-20/11.yml b/data/release-notes/enterprise-server/3-20/11.yml index 7d0b6a2c5ec6..427885434a0a 100644 --- a/data/release-notes/enterprise-server/3-20/11.yml +++ b/data/release-notes/enterprise-server/3-20/11.yml @@ -1,16 +1,17 @@ date: '2026-11-10' +intro: | + > [!WARNING] + > The Git security fixes previously listed for this release were incomplete. To receive these fixes, upgrade to {% data variables.product.prodname_ghe_server %} [3.20.12](#3.20.12) or a later patch release of 3.20. [Updated: 2026-11-17] sections: security_fixes: - | - **CRITICAL**: The Ruby Active Storage component was updated to address the file-read and remote code execution vulnerability [CVE-2026-66066](https://www.cve.org/CVERecord?id=CVE-2026-66066). + Packages have been updated to the latest security versions. - | - **HIGH**: Updated Git to include fixes that address [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469). + **CRITICAL**: The Ruby Active Storage component was updated to address the file-read and remote code execution vulnerability [CVE-2026-66066](https://www.cve.org/CVERecord?id=CVE-2026-66066). - | **MODERATE**: The Brotli compression library bundled with the GitHub application was updated to address [CVE-2026-111497](https://www.cve.org/CVERecord?id=CVE-2026-111497). - | Security updates have been applied to Go, OpenSSL, and Ubuntu system packages, including SQLite, across targeted main and bundled container images. These updates partially address [CVE-2026-111943](https://www.cve.org/CVERecord?id=CVE-2026-111943), [CVE-2025-15467](https://www.cve.org/CVERecord?id=CVE-2025-15467), and [CVE-2025-6965](https://www.cve.org/CVERecord?id=CVE-2025-6965); remaining components will be updated in a future release. - - | - Packages have been updated to the latest security versions. bugs: - | On large instances with a license for {% data variables.product.prodname_GHAS %}, applying a security configuration to many repositories could time out while checking license usage, leaving repositories that failed to enable. License usage checks are now faster under concurrent load. @@ -45,3 +46,6 @@ sections: When an administrator restores a backup with `ghe-restore` to a configured standalone instance with Replication Controller enabled, `ghe-restore` may skip storage data, such as avatars and attachments, and still complete successfully. The restore output includes `Warning: no routes found, skipping storage restore ...`. If you see this warning or storage data is missing after a restore, contact GitHub Support. - | After a restart, an instance can remain unavailable with "Starting..." displayed in the web interface. To recover, an administrator can connect using administrative SSH and rerun `ghe-config-apply`. This provides temporary recovery but does not prevent the issue from recurring. This issue is resolved in {% data variables.product.prodname_ghe_server %} 3.21.0 and later. + errata: + - | + The security fix note for [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469) has been removed because the Git security fixes were incomplete in this release. These fixes are included in {% data variables.product.prodname_ghe_server %} 3.20.12. [Updated: 2026-11-17] diff --git a/data/release-notes/enterprise-server/3-20/12.yml b/data/release-notes/enterprise-server/3-20/12.yml new file mode 100644 index 000000000000..f56d5af2d47b --- /dev/null +++ b/data/release-notes/enterprise-server/3-20/12.yml @@ -0,0 +1,147 @@ +date: '2026-11-17' +sections: + features: + - | + Administrators can use `--metrics-days N` with the support bundle utilities to collect metrics for a positive number of days independently of the log collection period. The `--metrics-only` option collects metrics, metrics-related logs, and bundle metadata without other diagnostics. By default, bundles include 10 days of metrics on standalone instances and clusters with up to 20 nodes. For larger clusters, the default decreases based on cluster size, to a minimum of three days. + security_fixes: + - | + Packages have been updated to the latest security versions. + - | + **HIGH:** Updated Git to include fixes that address [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469). This update was previously communicated in the release notes for the last patch release, but had not been fully resolved. + - | + **HIGH:** On a GitHub Enterprise Server instance in a cluster configuration, an attacker authenticated as a site administrator could execute arbitrary commands as the host `git` user on peer Git server nodes outside the pre-receive hook sandbox by supplying a crafted download URL for a custom pre-receive hook environment. GitHub has requested CVE ID [CVE-2026-112740](https://www.cve.org/cverecord?id=CVE-2026-112740) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** A site administrator could modify host files with root privileges and disclose private repository data due to improper path validation in pre-receive hook environment setup. To mitigate this issue, GitHub has strengthened path validation. GitHub has requested CVE ID [CVE-2026-116637](https://www.cve.org/cverecord?id=CVE-2026-116637) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An organization member could read and modify another user's private repositories due to improper SSH certificate identity validation. Exploitation required an instance using built-in authentication and a valid certificate issued to the attacker by a trusted SSH certificate authority. Access was limited to repositories within that authority's scope. GitHub has requested CVE ID [CVE-2026-116643](https://www.cve.org/cverecord?id=CVE-2026-116643) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An attacker with a session for a user whose numeric ID matched an enterprise's ID could generate private keys for its GitHub Apps by bypassing ownership checks. Repository access depended on the App's installations and permissions. On standard fresh installations, the collision involved the non-loginable ghost account; the reported reproduction required site administrator impersonation. GitHub has requested CVE ID [CVE-2026-110307](https://www.cve.org/cverecord?id=CVE-2026-110307) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH**: An attacker with permission to manage custom-property definitions and edit values could add an SSH key to another member's personal account through stored HTML in a single-select option. Exploitation required a definition manager in the same organization to attempt to remove the in-use option and save with a valid sudo session. GitHub has requested CVE ID [CVE-2026-131800](https://www.cve.org/cverecord?id=CVE-2026-131800) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/) + - | + **HIGH:** A user with repository write access could exploit server-side request forgery in Slack webhook validity checks, potentially leading to remote code execution on the appliance. Exploitation required Secret Protection with secret scanning and validity checks for partner patterns enabled. GitHub has requested CVE ID [CVE-2026-131820](https://www.cve.org/cverecord?id=CVE-2026-131820) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An OS command injection vulnerability in the GitHub Pages deployment process could allow arbitrary command execution as the GitHub Pages service user on a GitHub Enterprise Server instance. Exploitation required an authenticated user with permission to deploy a GitHub Pages site. GitHub has requested CVE ID [CVE-2026-130824](https://www.cve.org/cverecord?id=CVE-2026-130824) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **MEDIUM:** A user with write access could bypass branch protections through specially crafted input when resolving pull request conflicts. GitHub has requested CVE ID [CVE-2026-131647](https://www.cve.org/cverecord?id=CVE-2026-131647) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **LOW:** SAML authentication could match an incorrect account when NameID values differed only in case, accents, or trailing spaces. These mismatches now return a login-conflict error. + - | + **LOW:** Secret scanning validity checks for Google Cloud service account credentials were hardened to prevent HTTP redirects from sending validation requests to unintended destinations. Redirect responses leave the secret's validity as unknown. + - | + An attacker with write access to a repository with secret scanning enabled could execute arbitrary code on the instance by committing a crafted Google Cloud service account credential whose token URI pointed to an internal service on the instance. When GitHub Enterprise Server ran a validity check on the detected secret, it sent a request to the attacker-controlled URL. Validity check requests are now blocked from reaching internal services on the instance. This vulnerability was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + bugs: + - | + When users cloned or fetched a repository over SSH with compression enabled, the operation could stall until the connection timed out, especially for repositories with many references. + - | + Site administrators could not apply configuration changes when the backup CPU quota (`backup.cpu-quota`) was set to a valid percentage, such as `320%` or `12.5%`, because configuration validation rejected the percent sign. + - | + On instances configured with multiple trusted IP addresses or CIDR ranges for the Prometheus metrics endpoint, `ghe-config-check` reported validation errors for valid comma-separated lists. + - | + On instances with legacy metrics enabled, the weekly cleanup did not remove metrics files that had not been updated for more than 90 days. These stale files continued to consume space on the root disk. + - | + Site administrators could not apply configuration changes when the backup schedule (`backup.schedule`) was set to a valid cron schedule, such as `0 0 * * *`, because configuration validation rejected whitespace and `*`. + - | + Administrators restoring backups on standalone instances could find that Redis loaded existing append-only file (AOF) data instead of the restored backup. Redis restores could also fail for backups in RDB version 11 format. + - | + Site administrators could not apply configuration after upgrading an instance with a custom OpenTelemetry pipeline that referenced the built-in `otlp/ghes` receiver. + - | + Site administrators whose GitHub Actions storage settings included values such as "github" saw those values incorrectly redacted from hostnames and database names in support bundles, resulting in false diagnostic failures. + - | + After OpenTelemetry metrics were disabled, collectd could stop receiving application and service metrics until it was restarted. + - | + Administrators running `ghe-live-migrations` could see duplicate timestamps and stray formatting characters in the output. Sending SIGTERM to the utility did not always stop the underlying migration container cleanly, and the utility did not reject some invalid options. + - | + When an administrator changed `redis.io-threads` and applied the configuration, Redis on the primary could restart with the previous value, so the change did not take effect. + - | + After a site administrator applied a hotpatch, some files that the hotpatch installed, such as command-line utilities, could be missing from the instance. + - | + Upgrades and other operations that drain a node could take about 10 seconds longer than necessary because an internal service did not shut down promptly when signaled. + - | + After an upgrade, `ghe-config-apply` verified the same Docker images more than once, unnecessarily extending the configuration run. + - | + Site administrators could not trigger an Elasticsearch snapshot backup by setting `backup.es-backup-strategy` in `ghe-config`. + - | + On standalone or high-availability instances, repository data could be restored even when the corresponding repository metadata was missing from the database backup. Newly created repositories could then be preinitialized with that orphaned data. + - | + Administrators configuring repository caches on instances with GitHub Actions enabled saw a `mssql-server and mysql-server should have the same value` warning during replication setup or configuration runs. + - | + On instances with a large number of committers, enabling security features on a repository, including by applying a security configuration, could time out while calculating remaining GitHub Advanced Security licenses. This calculation is now skipped when enabling the repository uses no additional licenses. + - | + If calculating {% data variables.product.prodname_GH_advanced_security %} license usage took too long while an enterprise owner applied a security configuration, the confirmation dialog did not show how many repositories the configuration would apply to. + - | + When site administrators added or re-added a repository cache replica, repositories could remain unavailable on the cache, causing requests routed to it to return `Repository not found` errors until the background scheduler repaired them. + - | + Site administrators could encounter an error when viewing pending {% data variables.product.pat_v2 %} requests associated with a deleted token. + - | + Organizations could receive email notifications about pending {% data variables.product.pat_v2 %} requests after the associated tokens had been deleted. + - | + On instances running in private mode with subdomain isolation enabled, users could intermittently encounter broken avatar images and failures when loading content from the media, notebooks, and viewscreen subdomains. + - | + Secret scanning alerts for enterprise-level custom patterns did not include links to the custom pattern settings. + - | + For teams linked to SCIM external groups, the site administrator dashboard displayed a **Sync** action that called a service unavailable on GitHub Enterprise Server and a description that incorrectly indicated membership would reconcile automatically. Site administrators can use the **Reconcile** action to reconcile team membership. + - | + Background processes handling issue and project updates could consume increasing amounts of memory. + - | + Updates to organization rulesets targeting many repositories could time out. + - | + When GraphQL requests timed out, {% data variables.product.prodname_ghe_server %} could include all executed SQL queries in exception diagnostics, causing logs to grow enough to exhaust the root disk. These queries are no longer included in exception diagnostics. + - | + Users could be incorrectly classified as dormant despite recent activity, such as commenting on issues or pull requests. Last activity timestamps in user reports and the dormant users page could also be outdated. + - | + Users could not upload HTML and other supported file types as attachments to issues, pull requests, discussions, and comments in repositories. + - | + On instances with high push activity, users could experience delays before repository-related updates appeared in the UI. {% data variables.product.prodname_ghe_server %} processes repository push events with increased parallelism to reduce these delays. + changes: + - | + On AWS, the root volume for new instances launched from the GitHub Enterprise Server Amazon Machine Image (AMI) defaults to the gp3 volume type with 250 MiB/s throughput, instead of gp2. + - | + On clustered instances, the `ghe-cluster-balance` utility no longer displays an outdated disclaimer that described it as a preview feature and directed administrators to contact GitHub Support. The utility has been generally available and fully supported since {% data variables.product.prodname_ghe_server %} 3.9. For more information, see [AUTOTITLE](/admin/monitoring-and-managing-your-instance/configuring-clustering/rebalancing-cluster-workloads). + - | + Redis usage diagnostics use more responsive and reliable keyspace scans, including on larger installations. Collection is bounded by a timeout so diagnostics can continue when Redis scans take too long. + - | + The `ghe-backup-disk-extend` command-line utility is available to increase the attached disk volume size used by the backup service. + - | + Site administrators can run pre-upgrade live migrations in the background with `ghe-live-migrations --all --background` and check their status with `ghe-live-migrations --status`, which shows whether a migration is running and the latest lines of its log. When run from an interactive terminal, `ghe-live-migrations --all` prompts administrators to choose whether to run in the background. + - | + Site administrators can limit log collection when generating support bundles. The `--syslog-only` option collects syslog and MSSQL Server logs, while `--no-logs` skips application log collection. Both options retain metrics and selected diagnostics and cannot be used together. + - | + Administrators can configure the timeout for GitHub Actions queue-build API requests. The default is 10 seconds. To increase it to 30 seconds, run `ghe-config app.actions.queue-build-timeout 30s`, then `ghe-config-apply`. + - | + The bundled GitHub Actions runner has been updated to version 2.338.2. + - | + On instances where calculating GitHub Advanced Security license usage takes too long, applying a security configuration to repositories can fail. With help from GitHub Support, site administrators can skip license seat checks when security features are enabled. + known_issues: + - | + During an upgrade of GitHub Enterprise Server, custom firewall rules are removed. If you use custom firewall rules, you must reapply them after upgrading. + - | + If the root site administrator is locked out of the Management Console after failed login attempts, the account does not unlock automatically after the defined lockout time. Someone with administrative SSH access to the instance must unlock the account using the administrative shell. For more information, see [AUTOTITLE](/admin/administering-your-instance/administering-your-instance-from-the-web-ui/troubleshooting-access-to-the-management-console#unlocking-the-root-site-administrator-account). + - | + {% data reusables.release-notes.large-adoc-files-issue %} + - | + Admin stats REST API endpoints may time out on appliances with many users or repositories. Retrying the request until data is returned is advised. + - | + When following the steps for [Replacing the primary database node](/admin/monitoring-and-managing-your-instance/configuring-clustering/replacing-a-cluster-node#replacing-the-primary-database-node-mysql-or-mysql-and-mssql), running `ghe-cluster-config-apply` might fail with errors. If this occurs, re-running `ghe-cluster-config-apply` is expected to succeed. + - | + Running a config apply as part of the steps for [Replacing a node in an emergency](/admin/monitoring-and-managing-your-instance/configuring-clustering/replacing-a-cluster-node#replacing-a-node-in-an-emergency) may fail with errors if the node being replaced is still reachable. If this occurs, shut down the node and repeat the steps. + - | + When restoring data originally backed up from a 3.13 or greater appliance version, the Elasticsearch indices need to be reindexed before some of the data will show up. This happens via a scheduled job that runs on the first Saturday of each month. It can also be forced by running `/usr/local/share/enterprise/ghe-es-search-repair`. + - | + When initializing a new GHES cluster, nodes with the `consul-server` role should be added to the cluster before adding additional nodes. Adding all nodes simultaneously creates a race condition between nomad server registration and nomad client registration. + - | + In a cluster, the host running restore requires access to the storage nodes via their private IPs. + - | + After a restore, existing outside collaborators cannot be added to repositories in a new organization. This issue can be resolved by running `/usr/local/share/enterprise/ghe-es-search-repair` on the appliance. + - | + After a geo-replica is promoted to be a primary by running `ghe-repl-promote`, the actions workflow of a repository does not have any suggested workflows. + - | + When publishing npm packages in a workflow shortly after restoring from a backup to GitHub Enterprise Server, you may encounter a `401 Unauthorized` error from the GitHub Packages service. This has been confirmed on earlier releases and may also affect later supported versions. Applicability to current releases is still being confirmed. This can happen if the restore is from an N-1 or N-2 version and the workflow targets the npm endpoint on the backup instance. To avoid this issue, ensure the access token used for publishing is valid and includes the correct scopes for the GitHub Packages service. + - | + On instances with multiple Git storage nodes in a voting configuration, including cluster and geo-replication high availability topologies, upgrading may fail to correctly install Actions that ship with the new version. In some cases, previous versions of these Actions remain on the instance. To resolve this issue, run the following commands on the primary node: `ghe-config --unset 'app.actions.actions-repos-sha1sum'`, `ghe-config-apply`, and `/usr/local/share/enterprise/ghe-run-init-actions-graph`. + - | + When an administrator restores a backup with `ghe-restore` to a configured standalone instance with Replication Controller enabled, `ghe-restore` may skip storage data, such as avatars and attachments, and still complete successfully. The restore output includes `Warning: no routes found, skipping storage restore ...`. If you see this warning or storage data is missing after a restore, contact GitHub Support. + - | + After a restart, an instance can remain unavailable with "Starting..." displayed in the web interface. To recover, an administrator can connect using administrative SSH and rerun `ghe-config-apply`. This provides temporary recovery but does not prevent the issue from recurring. This issue is resolved in {% data variables.product.prodname_ghe_server %} 3.21.0 and later. + - | + `ghe-storage-extend` will fail for disks with a sector size greater than 1KiB, most notably with Azure Premium SSD v2 disks. This is due to an upstream bug with `resize2fs` when resizing a mounted disk. diff --git a/data/release-notes/enterprise-server/3-21/10.yml b/data/release-notes/enterprise-server/3-21/10.yml new file mode 100644 index 000000000000..c66dc79dfbca --- /dev/null +++ b/data/release-notes/enterprise-server/3-21/10.yml @@ -0,0 +1,165 @@ +date: '2026-11-17' +sections: + features: + - | + Administrators can use `--metrics-days N` with the support bundle utilities to collect metrics for a positive number of days independently of the log collection period. The `--metrics-only` option collects metrics, metrics-related logs, and bundle metadata without other diagnostics. By default, bundles include 10 days of metrics on standalone instances and clusters with up to 20 nodes. For larger clusters, the default decreases based on cluster size, to a minimum of three days. + security_fixes: + - | + Packages have been updated to the latest security versions. + - | + **HIGH:** Updated Git to include fixes that address [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469). This update was previously communicated in the release notes for the last patch release, but had not been fully resolved. + - | + **HIGH:** On a GitHub Enterprise Server instance in a cluster configuration, an attacker authenticated as a site administrator could execute arbitrary commands as the host `git` user on peer Git server nodes outside the pre-receive hook sandbox by supplying a crafted download URL for a custom pre-receive hook environment. GitHub has requested CVE ID [CVE-2026-112740](https://www.cve.org/cverecord?id=CVE-2026-112740) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** A site administrator could modify host files with root privileges and disclose private repository data due to improper path validation in pre-receive hook environment setup. To mitigate this issue, GitHub has strengthened path validation. GitHub has requested CVE ID [CVE-2026-116637](https://www.cve.org/cverecord?id=CVE-2026-116637) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An organization member could read and modify another user's private repositories due to improper SSH certificate identity validation. Exploitation required an instance using built-in authentication and a valid certificate issued to the attacker by a trusted SSH certificate authority. Access was limited to repositories within that authority's scope. GitHub has requested CVE ID [CVE-2026-116643](https://www.cve.org/cverecord?id=CVE-2026-116643) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An authenticated attacker could read private file paths and available source-code excerpts by supplying another repository's review-thread IDs through a pull request they could access. This authorization bypass required no victim interaction. GitHub has requested CVE ID [CVE-2026-111801](https://www.cve.org/cverecord?id=CVE-2026-111801) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An attacker with a session for a user whose numeric ID matched an enterprise's ID could generate private keys for its GitHub Apps by bypassing ownership checks. Repository access depended on the App's installations and permissions. On standard fresh installations, the collision involved the non-loginable ghost account; the reported reproduction required site administrator impersonation. GitHub has requested CVE ID [CVE-2026-110307](https://www.cve.org/cverecord?id=CVE-2026-110307) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH**: An attacker with permission to manage custom-property definitions and edit values could add an SSH key to another member's personal account through stored HTML in a single-select option. Exploitation required a definition manager in the same organization to attempt to remove the in-use option and save with a valid sudo session. GitHub has requested CVE ID [CVE-2026-131800](https://www.cve.org/cverecord?id=CVE-2026-131800) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/) + - | + **HIGH:** A user with repository write access could exploit server-side request forgery in Slack webhook validity checks, potentially leading to remote code execution on the appliance. Exploitation required Secret Protection with secret scanning and validity checks for partner patterns enabled. GitHub has requested CVE ID [CVE-2026-131820](https://www.cve.org/cverecord?id=CVE-2026-131820) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An OS command injection vulnerability in the GitHub Pages deployment process could allow arbitrary command execution as the GitHub Pages service user on a GitHub Enterprise Server instance. Exploitation required an authenticated user with permission to deploy a GitHub Pages site. GitHub has requested CVE ID [CVE-2026-130824](https://www.cve.org/cverecord?id=CVE-2026-130824) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **MEDIUM:** A user with write access could bypass branch protections through specially crafted input when resolving pull request conflicts. GitHub has requested CVE ID [CVE-2026-131647](https://www.cve.org/cverecord?id=CVE-2026-131647) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **LOW:** SAML authentication could match an incorrect account when NameID values differed only in case, accents, or trailing spaces. These mismatches now return a login-conflict error. + - | + **LOW:** Secret scanning validity checks for Google Cloud service account credentials were hardened to prevent HTTP redirects from sending validation requests to unintended destinations. Redirect responses leave the secret's validity as unknown. + - | + An attacker with write access to a repository with secret scanning enabled could execute arbitrary code on the instance by committing a crafted Google Cloud service account credential whose token URI pointed to an internal service on the instance. When GitHub Enterprise Server ran a validity check on the detected secret, it sent a request to the attacker-controlled URL. Validity check requests are now blocked from reaching internal services on the instance. This vulnerability was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + bugs: + - | + When users cloned or fetched a repository over SSH with compression enabled, the operation could stall until the connection timed out, especially for repositories with many references. + - | + Administrators could not apply SSH cipher configurations containing only CTR ciphers because `ghe-crypto` and the Management Console incorrectly rejected these lists. + - | + Site administrators could not apply configuration changes when the backup CPU quota (`backup.cpu-quota`) was set to a valid percentage, such as `320%` or `12.5%`, because configuration validation rejected the percent sign. + - | + On instances configured with multiple trusted IP addresses or CIDR ranges for the Prometheus metrics endpoint, `ghe-config-check` reported validation errors for valid comma-separated lists. + - | + On instances with legacy metrics enabled, the weekly cleanup did not remove metrics files that had not been updated for more than 90 days. These stale files continued to consume space on the root disk. + - | + Site administrators could not apply configuration changes when the backup schedule (`backup.schedule`) was set to a valid cron schedule, such as `0 0 * * *`, because configuration validation rejected whitespace and `*`. + - | + Administrators restoring backups on standalone instances could find that Redis loaded existing append-only file (AOF) data instead of the restored backup. Redis restores could also fail for backups in RDB version 11 format. + - | + Site administrators could not apply configuration after upgrading an instance with a custom OpenTelemetry pipeline that referenced the built-in `otlp/ghes` receiver. + - | + Site administrators whose GitHub Actions storage settings included values such as "github" saw those values incorrectly redacted from hostnames and database names in support bundles, resulting in false diagnostic failures. + - | + After OpenTelemetry metrics were disabled, collectd could stop receiving application and service metrics until it was restarted. + - | + Administrators running `ghe-live-migrations` could see duplicate timestamps and stray formatting characters in the output. Sending SIGTERM to the utility did not always stop the underlying migration container cleanly, and the utility did not reject some invalid options. + - | + When an administrator changed `redis.io-threads` and applied the configuration, Redis on the primary could restart with the previous value, so the change did not take effect. + - | + After a site administrator applied a hotpatch, some files that the hotpatch installed, such as command-line utilities, could be missing from the instance. + - | + On clusters with automatic MySQL failover enabled, upgrades could continue without successfully disabling failover, potentially causing an unexpected promotion of a MySQL replica. Upgrades stop before shutting down MySQL if failover suppression cannot be confirmed. + - | + Upgrades and other operations that drain a node could take about 10 seconds longer than necessary because an internal service did not shut down promptly when signaled. + - | + After an upgrade, `ghe-config-apply` verified the same Docker images more than once, unnecessarily extending the configuration run. + - | + Configured appliances reused as replicas in a high availability configuration could come online before replication started. + - | + On instances configured to forward logs over UDP, a journal message of approximately 64 KiB or larger could stop all subsequent log forwarding until `syslog-ng` was restarted. + - | + Site administrators could not trigger an Elasticsearch snapshot backup by setting `backup.es-backup-strategy` in `ghe-config`. + - | + On standalone or high-availability instances, repository data could be restored even when the corresponding repository metadata was missing from the database backup. Newly created repositories could then be preinitialized with that orphaned data. + - | + Administrators configuring repository caches on instances with GitHub Actions enabled saw a `mssql-server and mysql-server should have the same value` warning during replication setup or configuration runs. + - | + Users with qualifying issue or pull request activity did not see the "Recent activity" section on their personal dashboards. + - | + On instances with a large number of committers, enabling security features on a repository, including by applying a security configuration, could time out while calculating remaining GitHub Advanced Security licenses. This calculation is now skipped when enabling the repository uses no additional licenses. + - | + If calculating {% data variables.product.prodname_GH_advanced_security %} license usage took too long while an enterprise owner applied a security configuration, the confirmation dialog did not show how many repositories the configuration would apply to. + - | + When site administrators added or re-added a repository cache replica, repositories could remain unavailable on the cache, causing requests routed to it to return `Repository not found` errors until the background scheduler repaired them. + - | + Site administrators could encounter an error when viewing pending {% data variables.product.pat_v2 %} requests associated with a deleted token. + - | + Organizations could receive email notifications about pending {% data variables.product.pat_v2 %} requests after the associated tokens had been deleted. + - | + On instances running in private mode with subdomain isolation enabled, users could intermittently encounter broken avatar images and failures when loading content from the media, notebooks, and viewscreen subdomains. + - | + Users experienced failed Git pushes when secret scanning evaluated delegated bypass requests configured with an organization custom role. + - | + Secret scanning alerts for enterprise-level custom patterns did not include links to the custom pattern settings. + - | + For teams linked to SCIM external groups, the site administrator dashboard displayed a **Sync** action that called a service unavailable on GitHub Enterprise Server and a description that incorrectly indicated membership would reconcile automatically. Site administrators can use the **Reconcile** action to reconcile team membership. + - | + Background processes handling issue and project updates could consume increasing amounts of memory. + - | + Updates to organization rulesets targeting many repositories could time out. + - | + Users could be incorrectly classified as dormant despite recent activity, such as commenting on issues or pull requests. Last activity timestamps in user reports and the dormant users page could also be outdated. + - | + When GraphQL requests timed out, {% data variables.product.prodname_ghe_server %} could include all executed SQL queries in exception diagnostics, causing logs to grow enough to exhaust the root disk. These queries are no longer included in exception diagnostics. + - | + Users could not upload HTML and other supported file types as attachments to issues, pull requests, discussions, and comments in repositories. + - | + On instances with high push activity, users could experience delays before repository-related updates appeared in the UI. {% data variables.product.prodname_ghe_server %} processes repository push events with increased parallelism to reduce these delays. + changes: + - | + For site administrators viewing monitoring dashboards in the Management Console, Grafana tooltips sort values in descending order. + - | + On AWS, the root volume for new instances launched from the GitHub Enterprise Server Amazon Machine Image (AMI) defaults to the gp3 volume type with 250 MiB/s throughput, instead of gp2. + - | + On clustered instances, the `ghe-cluster-balance` utility no longer displays an outdated disclaimer that described it as a preview feature and directed administrators to contact GitHub Support. The utility has been generally available and fully supported since {% data variables.product.prodname_ghe_server %} 3.9. For more information, see [AUTOTITLE](/admin/monitoring-and-managing-your-instance/configuring-clustering/rebalancing-cluster-workloads). + - | + Redis usage diagnostics use more responsive and reliable keyspace scans, including on larger installations. Collection is bounded by a timeout so diagnostics can continue when Redis scans take too long. + - | + GitHub Enterprise Server saves system, usage, and performance data from each node to `/data/user/common/telemetry-snapshot` before and after upgrades and hotpatches. Uploading the data to GitHub Support requires administrator consent. If no consent choice is saved, interactive installations prompt administrators. To change the setting, run `ghe-config upgrade.telemetry-upload-consent true` to allow uploads, or run `ghe-config upgrade.telemetry-upload-consent false` to keep the data local. + - | + The `ghe-backup-disk-extend` command-line utility is available to increase the attached disk volume size used by the backup service. + - | + Site administrators can run pre-upgrade live migrations in the background with `ghe-live-migrations --all --background` and check their status with `ghe-live-migrations --status`, which shows whether a migration is running and the latest lines of its log. When run from an interactive terminal, `ghe-live-migrations --all` prompts administrators to choose whether to run in the background. + - | + Administrators can configure the timeout for GitHub Actions queue-build API requests. The default is 10 seconds. To increase it to 30 seconds, run `ghe-config app.actions.queue-build-timeout 30s`, then `ghe-config-apply`. + - | + Site administrators can limit log collection when generating support bundles. The `--syslog-only` option collects syslog and MSSQL Server logs, while `--no-logs` skips application log collection. Both options retain metrics and selected diagnostics and cannot be used together. + - | + The bundled GitHub Actions runner has been updated to version 2.338.2. + - | + On instances where calculating GitHub Advanced Security license usage takes too long, applying a security configuration to repositories can fail. With help from GitHub Support, site administrators can skip license seat checks when security features are enabled. + - | + For users who merge pull requests in repositories that automatically delete head branches, merges complete faster because GitHub Enterprise Server deletes the head branch in the background after the merge, rather than while the merge is in progress. The head branch may remain visible briefly after the pull request is merged. For more information, see [AUTOTITLE](/repositories/configuring-branches-and-merges-in-your-repository/configuring-pull-request-merges/managing-the-automatic-deletion-of-branches). + known_issues: + - | + During an upgrade of GitHub Enterprise Server, custom firewall rules are removed. If you use custom firewall rules, you must reapply them after upgrading. + - | + If the root site administrator is locked out of the Management Console after failed login attempts, the account does not unlock automatically after the defined lockout time. Someone with administrative SSH access to the instance must unlock the account using the administrative shell. For more information, see [AUTOTITLE](/admin/administering-your-instance/administering-your-instance-from-the-web-ui/troubleshooting-access-to-the-management-console#unlocking-the-root-site-administrator-account). + - | + {% data reusables.release-notes.large-adoc-files-issue %} + - | + Admin stats REST API endpoints may time out on appliances with many users or repositories. Retrying the request until data is returned is advised. + - | + When following the steps for [Replacing the primary database node](/admin/monitoring-and-managing-your-instance/configuring-clustering/replacing-a-cluster-node#replacing-the-primary-database-node-mysql-or-mysql-and-mssql), running `ghe-cluster-config-apply` might fail with errors. If this occurs, re-running `ghe-cluster-config-apply` is expected to succeed. + - | + Running a config apply as part of the steps for [Replacing a node in an emergency](/admin/monitoring-and-managing-your-instance/configuring-clustering/replacing-a-cluster-node#replacing-a-node-in-an-emergency) may fail with errors if the node being replaced is still reachable. If this occurs, shut down the node and repeat the steps. + - | + When restoring data originally backed up from a 3.13 or greater appliance version, the Elasticsearch indices need to be reindexed before some of the data will show up. This happens via a scheduled job that runs on the first Saturday of each month. It can also be forced by running `/usr/local/share/enterprise/ghe-es-search-repair`. + - | + When initializing a new GHES cluster, nodes with the `consul-server` role should be added to the cluster before adding additional nodes. Adding all nodes simultaneously creates a race condition between nomad server registration and nomad client registration. + - | + In a cluster, the host running restore requires access to the storage nodes via their private IPs. + - | + After a restore, existing outside collaborators cannot be added to repositories in a new organization. This issue can be resolved by running `/usr/local/share/enterprise/ghe-es-search-repair` on the appliance. + - | + After a geo-replica is promoted to be a primary by running `ghe-repl-promote`, the actions workflow of a repository does not have any suggested workflows. + - | + When publishing npm packages in a workflow shortly after restoring from a backup to GitHub Enterprise Server, you may encounter a `401 Unauthorized` error from the GitHub Packages service. This has been confirmed on earlier releases and may also affect later supported versions. Applicability to current releases is still being confirmed. This can happen if the restore is from an N-1 or N-2 version and the workflow targets the npm endpoint on the backup instance. To avoid this issue, ensure the access token used for publishing is valid and includes the correct scopes for the GitHub Packages service. + - | + On instances with multiple Git storage nodes in a voting configuration, including cluster and geo-replication high availability topologies, upgrading may fail to correctly install Actions that ship with the new version. In some cases, previous versions of these Actions remain on the instance. To resolve this issue, run the following commands on the primary node: `ghe-config --unset 'app.actions.actions-repos-sha1sum'`, `ghe-config-apply`, and `/usr/local/share/enterprise/ghe-run-init-actions-graph`. + - | + When an administrator restores a backup with `ghe-restore` to a configured standalone instance with Replication Controller enabled, `ghe-restore` may skip storage data, such as avatars and attachments, and still complete successfully. The restore output includes `Warning: no routes found, skipping storage restore ...`. If you see this warning or storage data is missing after a restore, contact GitHub Support. + - | + `ghe-storage-extend` will fail for disks with a sector size greater than 1KiB, most notably with Azure Premium SSD v2 disks. This is due to an upstream bug with `resize2fs` when resizing a mounted disk. diff --git a/data/release-notes/enterprise-server/3-21/9.yml b/data/release-notes/enterprise-server/3-21/9.yml index ef3521ff1c06..f757b7756927 100644 --- a/data/release-notes/enterprise-server/3-21/9.yml +++ b/data/release-notes/enterprise-server/3-21/9.yml @@ -1,16 +1,17 @@ date: '2026-11-10' +intro: | + > [!WARNING] + > The Git security fixes previously listed for this release were incomplete. To receive these fixes, upgrade to {% data variables.product.prodname_ghe_server %} [3.21.10](#3.21.10) or a later patch release of 3.21. [Updated: 2026-11-17] sections: security_fixes: - | - **CRITICAL**: The Ruby Active Storage component was updated to address the file-read and remote code execution vulnerability [CVE-2026-66066](https://www.cve.org/CVERecord?id=CVE-2026-66066). + Packages have been updated to the latest security versions. - | - **HIGH**: Updated Git to include fixes that address [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469). + **CRITICAL**: The Ruby Active Storage component was updated to address the file-read and remote code execution vulnerability [CVE-2026-66066](https://www.cve.org/CVERecord?id=CVE-2026-66066). - | **MODERATE**: The Brotli compression library bundled with the GitHub application was updated to address [CVE-2026-111497](https://www.cve.org/CVERecord?id=CVE-2026-111497). - | Security updates have been applied to Go, OpenSSL, and Ubuntu system packages, including SQLite, across targeted main and bundled container images. These updates partially address [CVE-2026-111943](https://www.cve.org/CVERecord?id=CVE-2026-111943), [CVE-2025-15467](https://www.cve.org/CVERecord?id=CVE-2025-15467), and [CVE-2025-6965](https://www.cve.org/CVERecord?id=CVE-2025-6965); remaining components will be updated in a future release. - - | - Packages have been updated to the latest security versions. bugs: - | On large instances with a license for {% data variables.product.prodname_GHAS %}, applying a security configuration to many repositories could time out while checking license usage, leaving repositories that failed to enable. License usage checks are now faster under concurrent load. @@ -47,3 +48,6 @@ sections: In some cases, pull requests using auto-merge or merge queue may not merge automatically until mergeability is recalculated. - | On a newly booted GitHub Enterprise Server instance, the merge box on a newly created pull request can stay on "Checking for the ability to merge automatically" and not show the merge status. If encountered, refreshing the page shows the correct merge status. + errata: + - | + The security fix note for [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469) has been removed because the Git security fixes were incomplete in this release. These fixes are included in {% data variables.product.prodname_ghe_server %} 3.21.10. [Updated: 2026-11-17] diff --git a/data/release-notes/enterprise-server/3-22/4.yml b/data/release-notes/enterprise-server/3-22/4.yml index 9001e52db9bb..63842716f28a 100644 --- a/data/release-notes/enterprise-server/3-22/4.yml +++ b/data/release-notes/enterprise-server/3-22/4.yml @@ -1,16 +1,17 @@ date: '2026-11-10' +intro: | + > [!WARNING] + > The Git security fixes previously listed for this release were incomplete. To receive these fixes, upgrade to {% data variables.product.prodname_ghe_server %} [3.22.5](#3.22.5) or a later patch release of 3.22. [Updated: 2026-11-17] sections: security_fixes: - | - **CRITICAL**: The Ruby Active Storage component was updated to address the file-read and remote code execution vulnerability [CVE-2026-66066](https://www.cve.org/CVERecord?id=CVE-2026-66066). + Packages have been updated to the latest security versions. - | - **HIGH**: Updated Git to include fixes that address [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469). + **CRITICAL**: The Ruby Active Storage component was updated to address the file-read and remote code execution vulnerability [CVE-2026-66066](https://www.cve.org/CVERecord?id=CVE-2026-66066). - | **MODERATE**: The Brotli compression library bundled with the GitHub application was updated to address [CVE-2026-111497](https://www.cve.org/CVERecord?id=CVE-2026-111497). - | Security updates have been applied to Go, OpenSSL, and Ubuntu system packages, including SQLite, across targeted main and bundled container images. These updates partially address [CVE-2026-111943](https://www.cve.org/CVERecord?id=CVE-2026-111943), [CVE-2025-15467](https://www.cve.org/CVERecord?id=CVE-2025-15467), and [CVE-2025-6965](https://www.cve.org/CVERecord?id=CVE-2025-6965); remaining components will be updated in a future release. - - | - Packages have been updated to the latest security versions. bugs: - | On large instances with a license for {% data variables.product.prodname_GHAS %}, applying a security configuration to many repositories could time out while checking license usage, leaving repositories that failed to enable. License usage checks are now faster under concurrent load. @@ -47,3 +48,6 @@ sections: In some cases, pull requests using auto-merge or merge queue may not merge automatically until mergeability is recalculated. - | On a newly booted GitHub Enterprise Server instance, the merge box on a newly created pull request can stay on "Checking for the ability to merge automatically" and not show the merge status. If encountered, refreshing the page shows the correct merge status. + errata: + - | + The security fix note for [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469) has been removed because the Git security fixes were incomplete in this release. These fixes are included in {% data variables.product.prodname_ghe_server %} 3.22.5. [Updated: 2026-11-17] diff --git a/data/release-notes/enterprise-server/3-22/5.yml b/data/release-notes/enterprise-server/3-22/5.yml new file mode 100644 index 000000000000..e77a4d721b6f --- /dev/null +++ b/data/release-notes/enterprise-server/3-22/5.yml @@ -0,0 +1,159 @@ +date: '2026-11-17' +sections: + features: + - | + Administrators can use `--metrics-days N` with the support bundle utilities to collect metrics for a positive number of days independently of the log collection period. The `--metrics-only` option collects metrics, metrics-related logs, and bundle metadata without other diagnostics. By default, bundles include 10 days of metrics on standalone instances and clusters with up to 20 nodes. For larger clusters, the default decreases based on cluster size, to a minimum of three days. + - | + Administrators can monitor backup health without SSH or Management Console access using new metrics for backup status, duration, snapshot size, and retention count. These metrics are sent to the appliance's existing metrics collector. + security_fixes: + - | + Packages have been updated to the latest security versions. + - | + **HIGH:** Updated Git to include fixes that address [CVE-2026-105437](https://www.cve.org/CVERecord?id=CVE-2026-105437), [CVE-2026-111760](https://www.cve.org/CVERecord?id=CVE-2026-111760), [CVE-2026-113408](https://www.cve.org/CVERecord?id=CVE-2026-113408), [CVE-2026-114466](https://www.cve.org/CVERecord?id=CVE-2026-114466), [CVE-2026-114467](https://www.cve.org/CVERecord?id=CVE-2026-114467), and [CVE-2026-114469](https://www.cve.org/CVERecord?id=CVE-2026-114469). This update was previously communicated in the release notes for the last patch release, but had not been fully resolved. + - | + **HIGH:** On a GitHub Enterprise Server instance in a cluster configuration, an attacker authenticated as a site administrator could execute arbitrary commands as the host `git` user on peer Git server nodes outside the pre-receive hook sandbox by supplying a crafted download URL for a custom pre-receive hook environment. GitHub has requested CVE ID [CVE-2026-112740](https://www.cve.org/cverecord?id=CVE-2026-112740) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** A site administrator could modify host files with root privileges and disclose private repository data due to improper path validation in pre-receive hook environment setup. To mitigate this issue, GitHub has strengthened path validation. GitHub has requested CVE ID [CVE-2026-116637](https://www.cve.org/cverecord?id=CVE-2026-116637) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An organization member could read and modify another user's private repositories due to improper SSH certificate identity validation. Exploitation required an instance using built-in authentication and a valid certificate issued to the attacker by a trusted SSH certificate authority. Access was limited to repositories within that authority's scope. GitHub has requested CVE ID [CVE-2026-116643](https://www.cve.org/cverecord?id=CVE-2026-116643) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An authenticated attacker could read private file paths and available source-code excerpts by supplying another repository's review-thread IDs through a pull request they could access. This authorization bypass required no victim interaction. GitHub has requested CVE ID [CVE-2026-111801](https://www.cve.org/cverecord?id=CVE-2026-111801) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An attacker with a session for a user whose numeric ID matched an enterprise's ID could generate private keys for its GitHub Apps by bypassing ownership checks. Repository access depended on the App's installations and permissions. On standard fresh installations, the collision involved the non-loginable ghost account; the reported reproduction required site administrator impersonation. GitHub has requested CVE ID [CVE-2026-110307](https://www.cve.org/cverecord?id=CVE-2026-110307) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH**: An attacker with permission to manage custom-property definitions and edit values could add an SSH key to another member's personal account through stored HTML in a single-select option. Exploitation required a definition manager in the same organization to attempt to remove the in-use option and save with a valid sudo session. GitHub has requested CVE ID [CVE-2026-131800](https://www.cve.org/cverecord?id=CVE-2026-131800) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/) + - | + **HIGH:** A user with repository write access could exploit server-side request forgery in Slack webhook validity checks, potentially leading to remote code execution on the appliance. Exploitation required Secret Protection with secret scanning and validity checks for partner patterns enabled. GitHub has requested CVE ID [CVE-2026-131820](https://www.cve.org/cverecord?id=CVE-2026-131820) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **HIGH:** An OS command injection vulnerability in the GitHub Pages deployment process could allow arbitrary command execution as the GitHub Pages service user on a GitHub Enterprise Server instance. Exploitation required an authenticated user with permission to deploy a GitHub Pages site. GitHub has requested CVE ID [CVE-2026-130824](https://www.cve.org/cverecord?id=CVE-2026-130824) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **MEDIUM:** A user with write access could bypass branch protections through specially crafted input when resolving pull request conflicts. GitHub has requested CVE ID [CVE-2026-131647](https://www.cve.org/cverecord?id=CVE-2026-131647) for this vulnerability, which was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + - | + **LOW:** SAML authentication could match an incorrect account when NameID values differed only in case, accents, or trailing spaces. These mismatches now return a login-conflict error. + - | + **LOW:** Secret scanning validity checks for Google Cloud service account credentials were hardened to prevent HTTP redirects from sending validation requests to unintended destinations. Redirect responses leave the secret's validity as unknown. + - | + An attacker with write access to a repository with secret scanning enabled could execute arbitrary code on the instance by committing a crafted Google Cloud service account credential whose token URI pointed to an internal service on the instance. When GitHub Enterprise Server ran a validity check on the detected secret, it sent a request to the attacker-controlled URL. Validity check requests are now blocked from reaching internal services on the instance. This vulnerability was reported via the [GitHub Bug Bounty program](https://bounty.github.com/). + bugs: + - | + When users cloned or fetched a repository over SSH with compression enabled, the operation could stall until the connection timed out, especially for repositories with many references. + - | + On clusters with automatic MySQL failover enabled, upgrades could continue without successfully disabling failover, potentially causing an unexpected promotion of a MySQL replica. Upgrades stop before shutting down MySQL if failover suppression cannot be confirmed. + - | + On an instance with repository replicas, pre-receive hook synchronization could fail to select a healthy replica to copy repository data from. This occurred because of a checksum compatibility issue between versions of the internal replication service, as well as incomplete checks of replica health indicators. + - | + Site administrators could not apply configuration changes when the backup CPU quota (`backup.cpu-quota`) was set to a valid percentage, such as `320%` or `12.5%`, because configuration validation rejected the percent sign. + - | + On instances configured with multiple trusted IP addresses or CIDR ranges for the Prometheus metrics endpoint, `ghe-config-check` reported validation errors for valid comma-separated lists. + - | + On instances with legacy metrics enabled, the weekly cleanup did not remove metrics files that had not been updated for more than 90 days. These stale files continued to consume space on the root disk. + - | + Site administrators could not apply configuration changes when the backup schedule (`backup.schedule`) was set to a valid cron schedule, such as `0 0 * * *`, because configuration validation rejected whitespace and `*`. + - | + Administrators restoring backups on standalone instances could find that Redis loaded existing append-only file (AOF) data instead of the restored backup. Redis restores could also fail for backups in RDB version 11 format. + - | + Site administrators could not apply configuration after upgrading an instance with a custom OpenTelemetry pipeline that referenced the built-in `otlp/ghes` receiver. + - | + Site administrators whose GitHub Actions storage settings included values such as "github" saw those values incorrectly redacted from hostnames and database names in support bundles, resulting in false diagnostic failures. + - | + After OpenTelemetry metrics were disabled, collectd could stop receiving application and service metrics until it was restarted. + - | + When an administrator changed `redis.io-threads` and applied the configuration, Redis on the primary could restart with the previous value, so the change did not take effect. + - | + After a site administrator applied a hotpatch, some files that the hotpatch installed, such as command-line utilities, could be missing from the instance. + - | + When an upgrade or hotpatch package failed signature verification, `ghe-upgrade` could exit without an error message. + - | + Upgrades and other operations that drain a node could take about 10 seconds longer than necessary because an internal service did not shut down promptly when signaled. + - | + After an upgrade, `ghe-config-apply` verified the same Docker images more than once, unnecessarily extending the configuration run. + - | + Configured appliances reused as replicas in a high availability configuration could come online before replication started. + - | + On instances configured to forward logs over UDP, a journal message of approximately 64 KiB or larger could stop all subsequent log forwarding until `syslog-ng` was restarted. + - | + After a restore, search fields could not display search results. + - | + Site administrators could not trigger an Elasticsearch snapshot backup by setting `backup.es-backup-strategy` in `ghe-config`. + - | + On standalone or high-availability instances, repository data could be restored even when the corresponding repository metadata was missing from the database backup. Newly created repositories could then be preinitialized with that orphaned data. + - | + Administrators configuring repository caches on instances with GitHub Actions enabled saw a `mssql-server and mysql-server should have the same value` warning during replication setup or configuration runs. + - | + Users experienced failed Git pushes when secret scanning evaluated delegated bypass requests configured with an organization custom role. + - | + On instances with a large number of committers, enabling security features on a repository, including by applying a security configuration, could time out while calculating remaining GitHub Advanced Security licenses. This calculation is now skipped when enabling the repository uses no additional licenses. + - | + If calculating {% data variables.product.prodname_GH_advanced_security %} license usage took too long while an enterprise owner applied a security configuration, the confirmation dialog did not show how many repositories the configuration would apply to. + - | + When site administrators added or re-added a repository cache replica, repositories could remain unavailable on the cache, causing requests routed to it to return `Repository not found` errors until the background scheduler repaired them. + - | + On the Dependabot alerts page, searching for an assignee by a name that contained certain Chinese characters, such as some characters in the GB18030 standard, returned an error instead of matching users. + - | + On instances running in private mode with subdomain isolation enabled, users could intermittently encounter broken avatar images and failures when loading content from the media, notebooks, and viewscreen subdomains. + - | + API requests to webhook endpoints for organizations, repositories, and integrations were incorrectly rate limited. The rate limiter has been removed. + - | + Secret scanning alerts for enterprise-level custom patterns did not include links to the custom pattern settings. + - | + For teams linked to SCIM external groups, the site administrator dashboard displayed a **Sync** action that called a service unavailable on GitHub Enterprise Server and a description that incorrectly indicated membership would reconcile automatically. Site administrators can use the **Reconcile** action to reconcile team membership. + - | + Background processes handling issue and project updates could consume increasing amounts of memory. + - | + Users could be incorrectly classified as dormant despite recent activity, such as commenting on issues or pull requests. Last activity timestamps in user reports and the dormant users page could also be outdated. + - | + When GraphQL requests timed out, {% data variables.product.prodname_ghe_server %} could include all executed SQL queries in exception diagnostics, causing logs to grow enough to exhaust the root disk. These queries are no longer included in exception diagnostics. + - | + Users could not upload HTML and other supported file types as attachments to issues, pull requests, discussions, and comments in repositories. + - | + On instances with high push activity, users could experience delays before repository-related updates appeared in the UI. {% data variables.product.prodname_ghe_server %} processes repository push events with increased parallelism to reduce these delays. + changes: + - | + For site administrators viewing monitoring dashboards in the Management Console, Grafana tooltips sort values in descending order. + - | + On AWS, the root volume for new instances launched from the GitHub Enterprise Server Amazon Machine Image (AMI) defaults to the gp3 volume type with 250 MiB/s throughput, instead of gp2. + - | + On clustered instances, the `ghe-cluster-balance` utility no longer displays an outdated disclaimer that described it as a preview feature and directed administrators to contact GitHub Support. The utility has been generally available and fully supported since {% data variables.product.prodname_ghe_server %} 3.9. For more information, see [AUTOTITLE](/admin/monitoring-and-managing-your-instance/configuring-clustering/rebalancing-cluster-workloads). + - | + Redis usage diagnostics use more responsive and reliable keyspace scans, including on larger installations. Collection is bounded by a timeout so diagnostics can continue when Redis scans take too long. + - | + GitHub Enterprise Server saves system, usage, and performance data from each node to `/data/user/common/telemetry-snapshot` before and after upgrades and hotpatches. Uploading the data to GitHub Support requires administrator consent. If no consent choice is saved, interactive installations prompt administrators. To change the setting, run `ghe-config upgrade.telemetry-upload-consent true` to allow uploads, or run `ghe-config upgrade.telemetry-upload-consent false` to keep the data local. + - | + The `ghe-backup-disk-extend` command-line utility is available to increase the attached disk volume size used by the backup service. + - | + Site administrators can limit log collection when generating support bundles. The `--syslog-only` option collects syslog and MSSQL Server logs, while `--no-logs` skips application log collection. Both options retain metrics and selected diagnostics and cannot be used together. + - | + Administrators can configure the timeout for GitHub Actions queue-build API requests. The default is 10 seconds. To increase it to 30 seconds, run `ghe-config app.actions.queue-build-timeout 30s`, then `ghe-config-apply`. + - | + The bundled GitHub Actions runner has been updated to version 2.338.2. + - | + On instances where calculating GitHub Advanced Security license usage takes too long, applying a security configuration to repositories can fail. With help from GitHub Support, site administrators can skip license seat checks when security features are enabled. + known_issues: + - | + During an upgrade of GitHub Enterprise Server, custom firewall rules are removed. If you use custom firewall rules, you must reapply them after upgrading. + - | + If the root site administrator is locked out of the Management Console after failed login attempts, the account does not unlock automatically after the defined lockout time. Someone with administrative SSH access to the instance must unlock the account using the administrative shell. For more information, see [AUTOTITLE](/admin/administering-your-instance/administering-your-instance-from-the-web-ui/troubleshooting-access-to-the-management-console#unlocking-the-root-site-administrator-account). + - | + {% data reusables.release-notes.large-adoc-files-issue %} + - | + Admin stats REST API endpoints may time out on appliances with many users or repositories. Retrying the request until data is returned is advised. + - | + When following the steps for [AUTOTITLE](/admin/monitoring-and-managing-your-instance/configuring-clustering/replacing-a-cluster-node#replacing-the-primary-database-node-mysql-or-mysql-and-mssql), running `ghe-cluster-config-apply` might fail with errors. If this occurs, re-running `ghe-cluster-config-apply` is expected to succeed. + - | + Running a config apply as part of the steps for [Replacing a node in an emergency](/admin/monitoring-and-managing-your-instance/configuring-clustering/replacing-a-cluster-node#replacing-a-node-in-an-emergency) may fail with errors if the node being replaced is still reachable. If this occurs, shut down the node and repeat the steps. + - | + When restoring data originally backed up from a 3.13 or greater appliance version, the Elasticsearch indices need to be reindexed before some of the data will show up. This happens via a scheduled job that runs on the first Saturday of each month. It can also be forced by running `/usr/local/share/enterprise/ghe-es-search-repair`. + - | + When initializing a new GHES cluster, nodes with the `consul-server` role should be added to the cluster before adding additional nodes. Adding all nodes simultaneously creates a race condition between nomad server registration and nomad client registration. + - | + In a cluster, the host running restore requires access to the storage nodes via their private IPs. + - | + After a restore, existing outside collaborators cannot be added to repositories in a new organization. This issue can be resolved by running `/usr/local/share/enterprise/ghe-es-search-repair` on the appliance. + - | + After a geo-replica is promoted to be a primary by running `ghe-repl-promote`, the actions workflow of a repository does not have any suggested workflows. + - | + When publishing npm packages in a workflow shortly after restoring from a backup to GitHub Enterprise Server, you may encounter a `401 Unauthorized` error from the GitHub Packages service. This has been confirmed on earlier releases and may also affect later supported versions. Applicability to current releases is still being confirmed. This can happen if the restore is from an N-1 or N-2 version and the workflow targets the npm endpoint on the backup instance. To avoid this issue, ensure the access token used for publishing is valid and includes the correct scopes for the GitHub Packages service. + - | + On instances with multiple Git storage nodes in a voting configuration, including cluster and geo-replication high availability topologies, upgrading may fail to correctly install Actions that ship with the new version. In some cases, previous versions of these Actions remain on the instance. To resolve this issue, run the following commands on the primary node: `ghe-config --unset 'app.actions.actions-repos-sha1sum'`, `ghe-config-apply`, and `/usr/local/share/enterprise/ghe-run-init-actions-graph`. + - | + `ghe-storage-extend` will fail for disks with a sector size greater than 1KiB, most notably with Azure Premium SSD v2 disks. This is due to an upstream bug with `resize2fs` when resizing a mounted disk. diff --git a/data/reusables/release-notes/2026-11-saml-issues-non-org-members-fixed.md b/data/reusables/release-notes/2026-11-saml-issues-non-org-members-fixed.md new file mode 100644 index 000000000000..1bb023ae449c --- /dev/null +++ b/data/reusables/release-notes/2026-11-saml-issues-non-org-members-fixed.md @@ -0,0 +1 @@ +On instances with SAML authentication enabled, signed-in users who were not members of the organization that owns a repository could not load Issues pages in public or internal repositories, even when they could view the repository's Code tab. diff --git a/src/github-apps/lib/config.json b/src/github-apps/lib/config.json index 7f9c4fb8937b..f453f5f2a529 100644 --- a/src/github-apps/lib/config.json +++ b/src/github-apps/lib/config.json @@ -60,5 +60,5 @@ "2022-11-28" ] }, - "sha": "35829a5b1d44a61da310ace729feacf5580ecd96" + "sha": "c9c95fe34ced05fe076001f5c55704742bc8e7f8" } \ No newline at end of file diff --git a/src/graphql/data/fpt/changelog.json b/src/graphql/data/fpt/changelog.json index 5959ebb25d96..0061a8d0dd82 100644 --- a/src/graphql/data/fpt/changelog.json +++ b/src/graphql/data/fpt/changelog.json @@ -1,4 +1,26 @@ [ + { + "schemaChanges": [ + { + "title": "The GraphQL schema includes these changes:", + "changes": [ + "

Enum value DYNAMIC was removed from enum CopilotCodeReviewEffort

", + "

Enum value ADAPTIVE was added to enum CopilotCodeReviewEffort

", + "

Enum value DYNAMIC was removed from enum CopilotCodeReviewEffortLevel

", + "

Enum value ADAPTIVE was added to enum CopilotCodeReviewEffortLevel

", + "

Enum value DYNAMIC was removed from enum OrganizationCopilotCodeReviewEffortSetting

", + "

Enum value ADAPTIVE was added to enum OrganizationCopilotCodeReviewEffortSetting

", + "

Enum value DYNAMIC was removed from enum RepositoryCopilotCodeReviewEffortSetting

", + "

Enum value ADAPTIVE was added to enum RepositoryCopilotCodeReviewEffortSetting

", + "

Field crossRepoContextEnabled was added to object type RepositoryCopilotCodeReviewSettings

", + "

Input field crossRepoContextEnabled of type Boolean was added to input object type UpdateRepositoryCopilotCodeReviewSettingsInput

" + ] + } + ], + "previewChanges": [], + "upcomingChanges": [], + "date": "2026-11-17" + }, { "schemaChanges": [ { diff --git a/src/graphql/data/fpt/schema-copilot.json b/src/graphql/data/fpt/schema-copilot.json index c3c28397057b..dc712ed541d4 100644 --- a/src/graphql/data/fpt/schema-copilot.json +++ b/src/graphql/data/fpt/schema-copilot.json @@ -428,6 +428,13 @@ "id": "copilotcoderevieweffortlevel", "href": "/graphql/reference/copilot#enum-copilotcoderevieweffortlevel" }, + { + "name": "crossRepoContextEnabled", + "description": "

Whether reviews can use public repository search.

", + "type": "Boolean!", + "id": "boolean", + "href": "/graphql/reference/other#scalar-boolean" + }, { "name": "customInstructionsEnabled", "description": "

Whether reviews use the repository's custom instructions.

", @@ -509,12 +516,12 @@ "isDeprecated": false, "values": [ { - "name": "BALANCED", - "description": "

Balanced effort and depth.

" + "name": "ADAPTIVE", + "description": "

Copilot picks the effort for each pull request.

" }, { - "name": "DYNAMIC", - "description": "

Copilot picks the effort for each pull request.

" + "name": "BALANCED", + "description": "

Balanced effort and depth.

" }, { "name": "LITE", @@ -535,12 +542,12 @@ "isDeprecated": false, "values": [ { - "name": "BALANCED", - "description": "

Balanced effort and depth.

" + "name": "ADAPTIVE", + "description": "

Copilot picks the effort for each pull request.

" }, { - "name": "DYNAMIC", - "description": "

Copilot picks the effort for each pull request.

" + "name": "BALANCED", + "description": "

Balanced effort and depth.

" }, { "name": "LITE", @@ -675,12 +682,12 @@ "isDeprecated": false, "values": [ { - "name": "BALANCED", - "description": "

Balanced effort and depth.

" + "name": "ADAPTIVE", + "description": "

Copilot picks the effort for each pull request.

" }, { - "name": "DYNAMIC", - "description": "

Copilot picks the effort for each pull request.

" + "name": "BALANCED", + "description": "

Balanced effort and depth.

" }, { "name": "INHERIT", @@ -769,12 +776,12 @@ "isDeprecated": false, "values": [ { - "name": "BALANCED", - "description": "

Balanced effort and depth.

" + "name": "ADAPTIVE", + "description": "

Copilot picks the effort for each pull request.

" }, { - "name": "DYNAMIC", - "description": "

Copilot picks the effort for each pull request.

" + "name": "BALANCED", + "description": "

Balanced effort and depth.

" }, { "name": "INHERIT", @@ -951,6 +958,13 @@ "id": "string", "href": "/graphql/reference/other#scalar-string" }, + { + "name": "crossRepoContextEnabled", + "description": "

Whether reviews can use public repository search. Can only be set when the\npublic repository search setting is available for the repository.

", + "type": "Boolean", + "id": "boolean", + "href": "/graphql/reference/other#scalar-boolean" + }, { "name": "customInstructionsEnabled", "description": "

Whether reviews use the repository's custom instructions.

", diff --git a/src/graphql/data/fpt/schema.docs.graphql b/src/graphql/data/fpt/schema.docs.graphql index 875cedd1fd87..afe6950c419a 100644 --- a/src/graphql/data/fpt/schema.docs.graphql +++ b/src/graphql/data/fpt/schema.docs.graphql @@ -7819,14 +7819,14 @@ The effort level for a Copilot code review. For more information, see [Review ef """ enum CopilotCodeReviewEffort @docsCategory(name: "copilot") { """ - Balanced effort and depth. + Copilot picks the effort for each pull request. """ - BALANCED + ADAPTIVE """ - Copilot picks the effort for each pull request. + Balanced effort and depth. """ - DYNAMIC + BALANCED """ Lower effort, faster review. @@ -7844,14 +7844,14 @@ A Copilot code review effort level. For more information, see [Review effort lev """ enum CopilotCodeReviewEffortLevel @docsCategory(name: "copilot") { """ - Balanced effort and depth. + Copilot picks the effort for each pull request. """ - BALANCED + ADAPTIVE """ - Copilot picks the effort for each pull request. + Balanced effort and depth. """ - DYNAMIC + BALANCED """ Lower effort, faster review. @@ -37769,14 +37769,14 @@ information, see [Review effort level](https://docs.github.com/en/copilot/concep """ enum OrganizationCopilotCodeReviewEffortSetting @docsCategory(name: "copilot") { """ - Balanced effort and depth. + Copilot picks the effort for each pull request. """ - BALANCED + ADAPTIVE """ - Copilot picks the effort for each pull request. + Balanced effort and depth. """ - DYNAMIC + BALANCED """ Clear the organization's value so repositories inherit the default. @@ -57595,14 +57595,14 @@ see [Review effort level](https://docs.github.com/en/copilot/concepts/agent-capa """ enum RepositoryCopilotCodeReviewEffortSetting @docsCategory(name: "copilot") { """ - Balanced effort and depth. + Copilot picks the effort for each pull request. """ - BALANCED + ADAPTIVE """ - Copilot picks the effort for each pull request. + Balanced effort and depth. """ - DYNAMIC + BALANCED """ Clear the repository's value so it inherits the effort from its organization, enterprise, or GitHub's default. @@ -57629,6 +57629,11 @@ type RepositoryCopilotCodeReviewSettings @docsCategory(name: "copilot") { """ availableReviewEfforts: [CopilotCodeReviewEffortLevel!]! + """ + Whether reviews can use public repository search. + """ + crossRepoContextEnabled: Boolean! + """ Whether reviews use the repository's custom instructions. """ @@ -72826,6 +72831,12 @@ input UpdateRepositoryCopilotCodeReviewSettingsInput { """ clientMutationId: String + """ + Whether reviews can use public repository search. Can only be set when the + public repository search setting is available for the repository. + """ + crossRepoContextEnabled: Boolean + """ Whether reviews use the repository's custom instructions. """ diff --git a/src/graphql/data/ghec/schema-copilot.json b/src/graphql/data/ghec/schema-copilot.json index c3c28397057b..dc712ed541d4 100644 --- a/src/graphql/data/ghec/schema-copilot.json +++ b/src/graphql/data/ghec/schema-copilot.json @@ -428,6 +428,13 @@ "id": "copilotcoderevieweffortlevel", "href": "/graphql/reference/copilot#enum-copilotcoderevieweffortlevel" }, + { + "name": "crossRepoContextEnabled", + "description": "

Whether reviews can use public repository search.

", + "type": "Boolean!", + "id": "boolean", + "href": "/graphql/reference/other#scalar-boolean" + }, { "name": "customInstructionsEnabled", "description": "

Whether reviews use the repository's custom instructions.

", @@ -509,12 +516,12 @@ "isDeprecated": false, "values": [ { - "name": "BALANCED", - "description": "

Balanced effort and depth.

" + "name": "ADAPTIVE", + "description": "

Copilot picks the effort for each pull request.

" }, { - "name": "DYNAMIC", - "description": "

Copilot picks the effort for each pull request.

" + "name": "BALANCED", + "description": "

Balanced effort and depth.

" }, { "name": "LITE", @@ -535,12 +542,12 @@ "isDeprecated": false, "values": [ { - "name": "BALANCED", - "description": "

Balanced effort and depth.

" + "name": "ADAPTIVE", + "description": "

Copilot picks the effort for each pull request.

" }, { - "name": "DYNAMIC", - "description": "

Copilot picks the effort for each pull request.

" + "name": "BALANCED", + "description": "

Balanced effort and depth.

" }, { "name": "LITE", @@ -675,12 +682,12 @@ "isDeprecated": false, "values": [ { - "name": "BALANCED", - "description": "

Balanced effort and depth.

" + "name": "ADAPTIVE", + "description": "

Copilot picks the effort for each pull request.

" }, { - "name": "DYNAMIC", - "description": "

Copilot picks the effort for each pull request.

" + "name": "BALANCED", + "description": "

Balanced effort and depth.

" }, { "name": "INHERIT", @@ -769,12 +776,12 @@ "isDeprecated": false, "values": [ { - "name": "BALANCED", - "description": "

Balanced effort and depth.

" + "name": "ADAPTIVE", + "description": "

Copilot picks the effort for each pull request.

" }, { - "name": "DYNAMIC", - "description": "

Copilot picks the effort for each pull request.

" + "name": "BALANCED", + "description": "

Balanced effort and depth.

" }, { "name": "INHERIT", @@ -951,6 +958,13 @@ "id": "string", "href": "/graphql/reference/other#scalar-string" }, + { + "name": "crossRepoContextEnabled", + "description": "

Whether reviews can use public repository search. Can only be set when the\npublic repository search setting is available for the repository.

", + "type": "Boolean", + "id": "boolean", + "href": "/graphql/reference/other#scalar-boolean" + }, { "name": "customInstructionsEnabled", "description": "

Whether reviews use the repository's custom instructions.

", diff --git a/src/graphql/data/ghec/schema.docs.graphql b/src/graphql/data/ghec/schema.docs.graphql index 875cedd1fd87..afe6950c419a 100644 --- a/src/graphql/data/ghec/schema.docs.graphql +++ b/src/graphql/data/ghec/schema.docs.graphql @@ -7819,14 +7819,14 @@ The effort level for a Copilot code review. For more information, see [Review ef """ enum CopilotCodeReviewEffort @docsCategory(name: "copilot") { """ - Balanced effort and depth. + Copilot picks the effort for each pull request. """ - BALANCED + ADAPTIVE """ - Copilot picks the effort for each pull request. + Balanced effort and depth. """ - DYNAMIC + BALANCED """ Lower effort, faster review. @@ -7844,14 +7844,14 @@ A Copilot code review effort level. For more information, see [Review effort lev """ enum CopilotCodeReviewEffortLevel @docsCategory(name: "copilot") { """ - Balanced effort and depth. + Copilot picks the effort for each pull request. """ - BALANCED + ADAPTIVE """ - Copilot picks the effort for each pull request. + Balanced effort and depth. """ - DYNAMIC + BALANCED """ Lower effort, faster review. @@ -37769,14 +37769,14 @@ information, see [Review effort level](https://docs.github.com/en/copilot/concep """ enum OrganizationCopilotCodeReviewEffortSetting @docsCategory(name: "copilot") { """ - Balanced effort and depth. + Copilot picks the effort for each pull request. """ - BALANCED + ADAPTIVE """ - Copilot picks the effort for each pull request. + Balanced effort and depth. """ - DYNAMIC + BALANCED """ Clear the organization's value so repositories inherit the default. @@ -57595,14 +57595,14 @@ see [Review effort level](https://docs.github.com/en/copilot/concepts/agent-capa """ enum RepositoryCopilotCodeReviewEffortSetting @docsCategory(name: "copilot") { """ - Balanced effort and depth. + Copilot picks the effort for each pull request. """ - BALANCED + ADAPTIVE """ - Copilot picks the effort for each pull request. + Balanced effort and depth. """ - DYNAMIC + BALANCED """ Clear the repository's value so it inherits the effort from its organization, enterprise, or GitHub's default. @@ -57629,6 +57629,11 @@ type RepositoryCopilotCodeReviewSettings @docsCategory(name: "copilot") { """ availableReviewEfforts: [CopilotCodeReviewEffortLevel!]! + """ + Whether reviews can use public repository search. + """ + crossRepoContextEnabled: Boolean! + """ Whether reviews use the repository's custom instructions. """ @@ -72826,6 +72831,12 @@ input UpdateRepositoryCopilotCodeReviewSettingsInput { """ clientMutationId: String + """ + Whether reviews can use public repository search. Can only be set when the + public repository search setting is available for the repository. + """ + crossRepoContextEnabled: Boolean + """ Whether reviews use the repository's custom instructions. """ diff --git a/src/rest/data/fpt-2022-11-28/copilot.json b/src/rest/data/fpt-2022-11-28/copilot.json index d2de17f24a2f..1e6e8c309570 100644 --- a/src/rest/data/fpt-2022-11-28/copilot.json +++ b/src/rest/data/fpt-2022-11-28/copilot.json @@ -1318,7 +1318,8 @@ ], "severity_filter": "no_filter", "custom_instructions_enabled": true, - "mcp_enabled": true + "mcp_enabled": true, + "cross_repo_context_enabled": true }, "schema": { "title": "Repository Copilot Code Review Settings", @@ -1369,6 +1370,10 @@ "mcp_enabled": { "type": "boolean", "description": "Whether reviews can use MCP servers." + }, + "cross_repo_context_enabled": { + "type": "boolean", + "description": "Whether reviews can use public repository search." } }, "required": [ @@ -1376,7 +1381,8 @@ "available_review_efforts", "severity_filter", "custom_instructions_enabled", - "mcp_enabled" + "mcp_enabled", + "cross_repo_context_enabled" ] } } @@ -1464,6 +1470,11 @@ "type": "boolean", "name": "mcp_enabled", "description": "

Whether reviews can use MCP servers.

" + }, + { + "type": "boolean", + "name": "cross_repo_context_enabled", + "description": "

Whether reviews can use public repository search. Can only be set when the public repository search setting is available for the repository.

" } ], "descriptionHTML": "

Updates the Copilot code review settings for a repository. These settings tune how Copilot code review\nreviews pull requests in the repository. They do not enable Copilot code review, and they can be stored\nwhile an enterprise or organization policy disables Copilot code review.

\n

Omitted settings are not changed. Set review_effort to inherit to clear the repository's review effort\nso it inherits the effort from the organization, the enterprise, or GitHub's default. The whole request is\nvalidated before any change is made, so invalid input does not change any setting. Changes use the same\naudit logging and telemetry as the repository settings page.

\n

Repository administrators can use this endpoint.\nOAuth app tokens and personal access tokens (classic) need the repo scope to use this endpoint.

", @@ -1497,7 +1508,8 @@ ], "severity_filter": "low", "custom_instructions_enabled": true, - "mcp_enabled": true + "mcp_enabled": true, + "cross_repo_context_enabled": true }, "schema": { "title": "Repository Copilot Code Review Settings", @@ -1548,6 +1560,10 @@ "mcp_enabled": { "type": "boolean", "description": "Whether reviews can use MCP servers." + }, + "cross_repo_context_enabled": { + "type": "boolean", + "description": "Whether reviews can use public repository search." } }, "required": [ @@ -1555,7 +1571,8 @@ "available_review_efforts", "severity_filter", "custom_instructions_enabled", - "mcp_enabled" + "mcp_enabled", + "cross_repo_context_enabled" ] } } diff --git a/src/rest/data/fpt-2026-03-10/copilot.json b/src/rest/data/fpt-2026-03-10/copilot.json index d0b7e8b6d513..872692ba0d2e 100644 --- a/src/rest/data/fpt-2026-03-10/copilot.json +++ b/src/rest/data/fpt-2026-03-10/copilot.json @@ -1318,7 +1318,8 @@ ], "severity_filter": "no_filter", "custom_instructions_enabled": true, - "mcp_enabled": true + "mcp_enabled": true, + "cross_repo_context_enabled": true }, "schema": { "title": "Repository Copilot Code Review Settings", @@ -1369,6 +1370,10 @@ "mcp_enabled": { "type": "boolean", "description": "Whether reviews can use MCP servers." + }, + "cross_repo_context_enabled": { + "type": "boolean", + "description": "Whether reviews can use public repository search." } }, "required": [ @@ -1376,7 +1381,8 @@ "available_review_efforts", "severity_filter", "custom_instructions_enabled", - "mcp_enabled" + "mcp_enabled", + "cross_repo_context_enabled" ] } } @@ -1464,6 +1470,11 @@ "type": "boolean", "name": "mcp_enabled", "description": "

Whether reviews can use MCP servers.

" + }, + { + "type": "boolean", + "name": "cross_repo_context_enabled", + "description": "

Whether reviews can use public repository search. Can only be set when the public repository search setting is available for the repository.

" } ], "descriptionHTML": "

Updates the Copilot code review settings for a repository. These settings tune how Copilot code review\nreviews pull requests in the repository. They do not enable Copilot code review, and they can be stored\nwhile an enterprise or organization policy disables Copilot code review.

\n

Omitted settings are not changed. Set review_effort to inherit to clear the repository's review effort\nso it inherits the effort from the organization, the enterprise, or GitHub's default. The whole request is\nvalidated before any change is made, so invalid input does not change any setting. Changes use the same\naudit logging and telemetry as the repository settings page.

\n

Repository administrators can use this endpoint.\nOAuth app tokens and personal access tokens (classic) need the repo scope to use this endpoint.

", @@ -1497,7 +1508,8 @@ ], "severity_filter": "low", "custom_instructions_enabled": true, - "mcp_enabled": true + "mcp_enabled": true, + "cross_repo_context_enabled": true }, "schema": { "title": "Repository Copilot Code Review Settings", @@ -1548,6 +1560,10 @@ "mcp_enabled": { "type": "boolean", "description": "Whether reviews can use MCP servers." + }, + "cross_repo_context_enabled": { + "type": "boolean", + "description": "Whether reviews can use public repository search." } }, "required": [ @@ -1555,7 +1571,8 @@ "available_review_efforts", "severity_filter", "custom_instructions_enabled", - "mcp_enabled" + "mcp_enabled", + "cross_repo_context_enabled" ] } } diff --git a/src/rest/data/ghec-2022-11-28/copilot.json b/src/rest/data/ghec-2022-11-28/copilot.json index 7548edcb925c..68761e07d3a8 100644 --- a/src/rest/data/ghec-2022-11-28/copilot.json +++ b/src/rest/data/ghec-2022-11-28/copilot.json @@ -1318,7 +1318,8 @@ ], "severity_filter": "no_filter", "custom_instructions_enabled": true, - "mcp_enabled": true + "mcp_enabled": true, + "cross_repo_context_enabled": true }, "schema": { "title": "Repository Copilot Code Review Settings", @@ -1369,6 +1370,10 @@ "mcp_enabled": { "type": "boolean", "description": "Whether reviews can use MCP servers." + }, + "cross_repo_context_enabled": { + "type": "boolean", + "description": "Whether reviews can use public repository search." } }, "required": [ @@ -1376,7 +1381,8 @@ "available_review_efforts", "severity_filter", "custom_instructions_enabled", - "mcp_enabled" + "mcp_enabled", + "cross_repo_context_enabled" ] } } @@ -1464,6 +1470,11 @@ "type": "boolean", "name": "mcp_enabled", "description": "

Whether reviews can use MCP servers.

" + }, + { + "type": "boolean", + "name": "cross_repo_context_enabled", + "description": "

Whether reviews can use public repository search. Can only be set when the public repository search setting is available for the repository.

" } ], "descriptionHTML": "

Updates the Copilot code review settings for a repository. These settings tune how Copilot code review\nreviews pull requests in the repository. They do not enable Copilot code review, and they can be stored\nwhile an enterprise or organization policy disables Copilot code review.

\n

Omitted settings are not changed. Set review_effort to inherit to clear the repository's review effort\nso it inherits the effort from the organization, the enterprise, or GitHub's default. The whole request is\nvalidated before any change is made, so invalid input does not change any setting. Changes use the same\naudit logging and telemetry as the repository settings page.

\n

Repository administrators can use this endpoint.\nOAuth app tokens and personal access tokens (classic) need the repo scope to use this endpoint.

", @@ -1497,7 +1508,8 @@ ], "severity_filter": "low", "custom_instructions_enabled": true, - "mcp_enabled": true + "mcp_enabled": true, + "cross_repo_context_enabled": true }, "schema": { "title": "Repository Copilot Code Review Settings", @@ -1548,6 +1560,10 @@ "mcp_enabled": { "type": "boolean", "description": "Whether reviews can use MCP servers." + }, + "cross_repo_context_enabled": { + "type": "boolean", + "description": "Whether reviews can use public repository search." } }, "required": [ @@ -1555,7 +1571,8 @@ "available_review_efforts", "severity_filter", "custom_instructions_enabled", - "mcp_enabled" + "mcp_enabled", + "cross_repo_context_enabled" ] } } diff --git a/src/rest/data/ghec-2026-03-10/copilot.json b/src/rest/data/ghec-2026-03-10/copilot.json index bd6d74f0bdd4..dedf9e0a33ff 100644 --- a/src/rest/data/ghec-2026-03-10/copilot.json +++ b/src/rest/data/ghec-2026-03-10/copilot.json @@ -1318,7 +1318,8 @@ ], "severity_filter": "no_filter", "custom_instructions_enabled": true, - "mcp_enabled": true + "mcp_enabled": true, + "cross_repo_context_enabled": true }, "schema": { "title": "Repository Copilot Code Review Settings", @@ -1369,6 +1370,10 @@ "mcp_enabled": { "type": "boolean", "description": "Whether reviews can use MCP servers." + }, + "cross_repo_context_enabled": { + "type": "boolean", + "description": "Whether reviews can use public repository search." } }, "required": [ @@ -1376,7 +1381,8 @@ "available_review_efforts", "severity_filter", "custom_instructions_enabled", - "mcp_enabled" + "mcp_enabled", + "cross_repo_context_enabled" ] } } @@ -1464,6 +1470,11 @@ "type": "boolean", "name": "mcp_enabled", "description": "

Whether reviews can use MCP servers.

" + }, + { + "type": "boolean", + "name": "cross_repo_context_enabled", + "description": "

Whether reviews can use public repository search. Can only be set when the public repository search setting is available for the repository.

" } ], "descriptionHTML": "

Updates the Copilot code review settings for a repository. These settings tune how Copilot code review\nreviews pull requests in the repository. They do not enable Copilot code review, and they can be stored\nwhile an enterprise or organization policy disables Copilot code review.

\n

Omitted settings are not changed. Set review_effort to inherit to clear the repository's review effort\nso it inherits the effort from the organization, the enterprise, or GitHub's default. The whole request is\nvalidated before any change is made, so invalid input does not change any setting. Changes use the same\naudit logging and telemetry as the repository settings page.

\n

Repository administrators can use this endpoint.\nOAuth app tokens and personal access tokens (classic) need the repo scope to use this endpoint.

", @@ -1497,7 +1508,8 @@ ], "severity_filter": "low", "custom_instructions_enabled": true, - "mcp_enabled": true + "mcp_enabled": true, + "cross_repo_context_enabled": true }, "schema": { "title": "Repository Copilot Code Review Settings", @@ -1548,6 +1560,10 @@ "mcp_enabled": { "type": "boolean", "description": "Whether reviews can use MCP servers." + }, + "cross_repo_context_enabled": { + "type": "boolean", + "description": "Whether reviews can use public repository search." } }, "required": [ @@ -1555,7 +1571,8 @@ "available_review_efforts", "severity_filter", "custom_instructions_enabled", - "mcp_enabled" + "mcp_enabled", + "cross_repo_context_enabled" ] } } diff --git a/src/rest/lib/config.json b/src/rest/lib/config.json index 20ea62b6b4f4..c86b9698a8e1 100644 --- a/src/rest/lib/config.json +++ b/src/rest/lib/config.json @@ -45,5 +45,5 @@ ] } }, - "sha": "35829a5b1d44a61da310ace729feacf5580ecd96" + "sha": "c9c95fe34ced05fe076001f5c55704742bc8e7f8" } \ No newline at end of file diff --git a/src/webhooks/lib/config.json b/src/webhooks/lib/config.json index 73c9b650db22..c3d819aa4279 100644 --- a/src/webhooks/lib/config.json +++ b/src/webhooks/lib/config.json @@ -1,3 +1,3 @@ { - "sha": "35829a5b1d44a61da310ace729feacf5580ecd96" + "sha": "c9c95fe34ced05fe076001f5c55704742bc8e7f8" } \ No newline at end of file