I have been using XFS with some success on Linux NFS fileservers. In the past month, one server has been subjected to an extremely heavy load - lots of IO, many files created/read/deleted, full files
Can you run xfs_repair -n on the device and send us the output. It appears you have hit the now infamous corruption. Attachment: signature.asc Description: This is a digitally signed message part
After the last crash, I decided to run xfs_repair on all the XFS filesystems. There was indeed corruption - about 20 files ended up in lost+found which I subsequently removed. Do you want me to re-ru
you've hits a few different things... at least this one: Jan 6 16:00:05 foobar kernel: Filesystem "md(9,0)": xfs_log_write: reservation ran out. Need to up reservation can be solved by mounting with
As I mentioned, I had already run xfs_repair on these filesystems. I had an opportunity today to re-run xfs_repair and sure enough, there were inconsistencies needing repair. I've attached the output