On 20/06/15 04:24 PM, PaulNM wrote:
On 06/20/2015 06:50 AM, Gary Dale wrote:I don't think this helps, but a RAID6 array with one failed drive is the same as RAID5 for all practical purposes. With 8 drives, I'd be very leery of converting from RAID6 to RAID5. The risk of a second drive failing while re-adding the first is not negligible. I think the backup drive allows for the reshape to be interrupted (most mdadm reshapes can be) but I wouldn't want to bet my data on. Since you have already done that, backup everything you consider essential then test whether it can be interrupted. On another machine copy your bootflash onto a USB external drive and switch it in. If it works, you're set.I suspected as much. I'm really curious if the constant use of the backup file is normal, or if anyone knows of better documentation on how it's used.Does anyone have suggestions on the best way to interrupt the reshape? Supposedly it can survive clean shutdowns, but I'm also seeing references saying the array should be stopped first. Since root is on the array, I'm concerned the system will just hang and not actually shutdown.The system in question is actually one of two similar systems, and is used as a backup of the other. There's some historical data that could be lost, but it wouldn't be the end of the world. I'd really rather avoid having to get it back up again and spend a week recopying everything over. :(Yeah, I mostly agree with a partially degraded 6 array practically being a 5. I'm not thrilled with 5 either, we will be going back to 6 eventually. There are a few differences that make the conversion worth it for me, though. The main one being that sometimes the system randomly wouldn't boot up if the array is degraded, a major issue since I work on it remotely. It could be quite a while before we add drives. I need to do an upgrade to Jessie and convert a large ext4 fs to btrfs, so I need reliable boots. But who doesn't? ;)- PaulNM
A normal shutdown will stop the array before the machine powers off.The backup file is quite likely being continually updated to reflect the current state of the reshape. That's the purpose of having the file.
You can try the article at http://serverfault.com/questions/688207/how-to-auto-start-degraded-software-raid1-under-debian-8-0-0-on-boot re. booting with a degraded array. I haven't tried it myself but it looks like it should work.