I had an interesting experience today while rebooting my media servers before a planned upgrade. I have 6 media servers both CPS-UCS-2RU-K9
and CPS-MSP-2RU-K9. I went to reboot them remotely (one by one, not all at the same time) but they would not come back. I had to go to the server
and physically power it off and turn it back on again, then it would boot up. Looking at the CLI after the 2nd one failed it was getting stuck after
starting the event engine daemon, throwing up java error and on the other one I was getting a "udevd worker: failed while handling
devices/platform/microcode/firmware/microcode". This was happening on both RH 5.8 and 6.4. it only did it on 3 of the 6 servers. Anyone else have
trouble rebooting their servers? It scares me because I'm about to upgrade to 7.7 from 7.6.
Wow... Nope. Can't say I've seen the same.
However, I don't believe the CPS-MSP chassis are supported anymore :-/ Any correlation to the errors you were seeing regarding the microcode? Was it only being thrown on the MSP chassis?
All the best.
I just looked at the release notes, the CPS-MSP's are fine, it was the CIVS chassis that isnt supported anymore. They haven't been restarted in 2+ years. Do you know if when the servers are upgrading do they do a system restart, or just restart services?
My bad. I must have been thinking about the VSM 7.7 Recovery Media then. You are correct.
Also, I don't *think* the servers go for a full reboot unless some very specific network stack information or clock related settings get changed (sometimes with no advanced warning).
Come to think of it, I don't believe the whole server actually goes for a reboot during an upgrade, but I *do* plan on there being a box reboot if *anything* 'at the VSMC/CDAF level is modified'. Nameservers, NTP, timezone, etc...