|
fsck - file system errors question
Hello,
Story of a server repair, and a question:
- Ubuntu 14.04
- virtual machines at a cloud provider
Many of our servers either crashed, or had their disks go into read-only mode after a SAN (storage area network) outage. Upon reboot, fsck wanted to fix the errors, but the KVM console wasn't working very well, so then rebooted into Rescue Mode. Ran fsck.
Many errors were repaired such as
Inode 1313966 was part of the orphaned inode list. FIXED.
Inode 1313968 was part of the orphaned inode list. FIXED.
and
Free blocks count wrong for group #6 (21577, counted=21592).
Fix<y>? yes
This is happened on a number of servers, as above.
On one of the servers ( let's call it serverX ) I was experimenting and trying to fix it in the KVM console only. KVM Keyboard input wasn't working well, I rebooted it a few times. Stopped and Started. Stopped and Started. Finally, going back to Rescue Mode. Run fsck.
fsck finds only one single error now, that Mount Time is in the future. It fixes this one error.
QUESTION:
Why did ServerX not have as many errors?
More specifically: Could repeatedly stopping and starting, rebooting, rebooting, a server somehow trick fsck into thinking that errors have disappeared, when in fact they are still present? For example, after a server were shutdown, the filesystem is mistakenly marked as clean. Then subsequent attempts at fsck will fail to fix the disk.
Or, it was just a lucky coincidence, and this server didn't have errors.
|
