Skip to content

Replace switch backend variables with tagged deployment enum - #11286

Open
sion42x wants to merge 1 commit into
skaram/detect-softnpufrom
skaram/deployment-config
Open

Replace switch backend variables with tagged deployment enum#11286
sion42x wants to merge 1 commit into
skaram/detect-softnpufrom
skaram/deployment-config

Conversation

@sion42x

@sion42x sion42x commented Sep 12, 2026

Copy link
Copy Markdown
Contributor

Changes based on feedback in #11237. This method drastically simplifies the tracking of switch backend by sled-agent and improves downstream understanding by naming the types of configs. It also means no startup probe is run when production option (tofino) is chosen which is more in line with the current method. Current checked in configs are production, virtual, standalone, and custom.

Changes based on feedback in #11237. This
method drastically simplifies the tracking of switch backend
by sled-agent and improves downstream understanding by naming the
types of configs. It also means no startup probe is run when
`production` option (tofino) is chosen which is more in line with
the current method. Current checked in configs are `production`,
`virtual`, `standalone`, and `custom`.
@sion42x
sion42x requested a review from jmpesp September 12, 2026 14:23
@sion42x

sion42x commented Sep 12, 2026

Copy link
Copy Markdown
Contributor Author

@sion42x

sion42x commented Sep 12, 2026

Copy link
Copy Markdown
Contributor Author

Succeeded on Madrid + voxel. On Madrid:

[madrid]root@oxz_switch0:~# pilot host exec -c 'svcs -H -o state oxide/sled-agent; grep -c "kind = .production" /opt/oxide/sled-agent/pkg/config.toml' 14-17
14  BRM42220081        ok: online
1
15  BRM42220046        ok: online
1
16  BRM42220007        ok: online
1
17  BRM42220004        ok: online
1
[madrid]root@oxz_switch0:~# pilot host exec -c 'zoneadm list | grep -c oxz_switch; svcs -H -o state pumpkind 2>/dev/null' 14-17
14  BRM42220081        ok: 1
online
15  BRM42220046        ok: 0
online
16  BRM42220007        ok: 1
online
17  BRM42220004        ok: 0
online
[madrid]root@oxz_switch0:~# svcs -H -o state dendrite; svcprop -p config/mgmt dendrite 2>/dev/null; ls -l /dev/tofino/1
online
crw-------   1 root     sys      143,  1 Dec 28  1986 /dev/tofino/1
[madrid]root@oxz_switch0:~# svcprop -p config/board_rev dendrite
b
[madrid]root@oxz_switch0:~# pilot host exec -c 'grep -h -E "Found tofino node|Updated tofino|tofino device found" \
> /var/svc/log/oxide-sled-agent:default.log /var/svc/log/oxide-tofino-monitor:default.log \
> 2>/dev/null | cut -c1-150 | head -3' 14-17
14  BRM42220081        ok: {"msg":"Updated tofino","v":0,"name":"SledAgent","level":30,"time":"1986-12-28T00:00:45.226661561Z","hostname":"BRM42220081","pid":674,"component":"Ha
Dec 28 00:00:45.282 INFO tofino device found, unit: tofino-monitor
15  BRM42220046        ok: 
16  BRM42220007        ok: {"msg":"Updated tofino","v":0,"name":"SledAgent","level":30,"time":"1986-12-28T00:00:45.158379905Z","hostname":"BRM42220007","pid":674,"component":"Ha
Dec 28 00:00:45.208 INFO tofino device found, unit: tofino-monitor
17  BRM42220004        ok: 
[madrid]root@oxz_switch0:~# 

@sion42x

sion42x commented Sep 12, 2026

Copy link
Copy Markdown
Contributor Author

On voxel:

root@helios1:~/oxide/omicron# for i in g0 g1 g2; do echo "== $i"; voxel host exec -c '
svcs -H -o state oxide/sled-agent
grep -c "kind = .virtual" /opt/oxide/sled-agent/pkg/config.toml
grep "voxel role" /opt/cargo-bay/sled-config.toml
grep -h -E "found SoftNPU 9p device|probe failed|busy|skipping pumpkind|SledModeConfig|unknown field" \
  /var/svc/log/oxide-sled-agent:default.log | cut -c1-140
zoneadm list | grep -c oxz_switch' $i 2>&1 | grep -v "INFO set propolis"; done
== g0
online
2
# voxel role: scrimlet
{"msg":"not an Oxide sled; skipping pumpkind","v":0,"name":"SledAgent","level":30,"time":"2026-09-12T18:19:13.579336523Z","hostname":"g0","p
{"msg":"found SoftNPU 9p device","v":0,"name":"SledAgent","level":30,"time":"2026-09-12T18:19:14.575353458Z","hostname":"g0","pid":776,"comp
1
== g1
online
2
# voxel role: gimlet
{"msg":"not an Oxide sled; skipping pumpkind","v":0,"name":"SledAgent","level":30,"time":"2026-09-12T18:19:15.771940262Z","hostname":"g1","p
0
== g2
online
2
# voxel role: scrimlet
{"msg":"not an Oxide sled; skipping pumpkind","v":0,"name":"SledAgent","level":30,"time":"2026-09-12T18:19:19.654551391Z","hostname":"g2","p
{"msg":"found SoftNPU 9p device","v":0,"name":"SledAgent","level":30,"time":"2026-09-12T18:19:20.374239364Z","hostname":"g2","pid":789,"comp
1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant