It would appear that RedHat started to support and actively recommend XFS beginning somewhere RHEL 5.3.
This would mean being a lot safer from all ext3 fsck related pain and also gaining a good bit of extra performance. also no more disgusting crashes during online resize.
Right now I’m thinking almost hourly how to:
- get our customer’s support for the move (less downtime and less risk of corruption sometimes just isn’t enough)
- go about migrating roundabout 100TB without too much application downtime.
Currently these are both very time intensive issues, I guess we lose about 300 hours of uptime over all server per year to various ext3 topics, and it would take at least 30 hours just to get the approval and plan a migration – and then it needs to be done, too.
So the seemingly “best” way to go would be to do it partly via the systems lifecycle. But I beg to differ, after all we STILL have some ext3 based clusters that give us a **** full of pain and noone cares to change it.
Time to go angry mode, stop suffering from design errors and push for a change for everyones better?
Ah well, I got ONE day left till vacation.
May(be) there be upgrades.