Open augustuswm opened 2 weeks ago
I tried updating Sled 13 and ran in to the same error.
Edit: wrong issue
Hubris dumps (and mgs logs) from before and after the SP update attempt on Sled 13 are available at /staff/rack3/BRM42220064/2024-10-18
This should be fixed by https://github.com/oxidecomputer/hubris/pull/1905 . It applies to all SPs.
Ignition reset does not seem to reliably clear the error. Ignition off, waiting until it is confirmed off and then ignition on worked reliably to update dogfood (which also reproduced the error). This should go in the release notes for R12.
I will also update with the associated omicron commit/TUF repo which should have the fix.
When attempting to update the colo rack from R11 rc0 to R11 rc1 we encountered a failure during the first step of the update (sled 14 / switch 0).
The update was initiated with
wicket16 rack-update start --sled 14 --switch 0 --psc 0 --force-update-rot --force-update-sp --color always
and after the ROT update update succeeded it reported a failure in updating the SP:I captured a hubris dump from the SP and gathered logs from the sled 16 switch zone. The dump is available in
/staff/rack3/BRM42220013/2024-10-17
and the logs are as follows:Wicket reports the following for the sled SP:
Update attempts after both clearing the error and resetting via ignition resulted in the same failure.