In T297265, the host randomly freezes with kernel 5.10.46 which was downgraded from 5.10.70 preemptively to mitigate a bug affecting the mx servers.
We'd like to see if a firmware upgrade helps mitigate the conntrack bug. Graphite1004 is currently the backup host.
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | fgiunchedi | T297265 graphite1004 freezing | |||
| Resolved | • Cmjohnson | T297433 Upgrade firmware on graphite1004 if upgrade available. |
@colewhite we can update the f/w but this will require the server to be out of production for about 30 minutes. We can do this almost anytime, let me know when you would like to proceed.
I think it just needs to be downtimed in icinga for the maintenance window. Being the backup host, I think you can proceed when you're ready.
graphite1004 f/w has been updated except the idrac. its version is too old and the system will not allow an update to the current or previous version. Hopefully, this helps, if not please re-open and I can investigate for any h/w issues.