Node disk stuck on journal recovery

I got a node, which turned itself off Saturday. There was some trouble with the file system and since this was the second time in a month, I decided to recover it to another disk. This went without any failures, but the disk turned out not mountable.

So, I ran e2fsck -vyf {partition} which however is running over a day now, still recovering the journal. Iostat tells there’s is some disk activity in range of KBs per second.

I don’t believe this is gonna work, so I’m essentially looking for an option to clear the journal. Like the -L for xfs_repair.

Any ideas?

Since you use xfs, you should run a xfs_repair instead of fsck.

It’s an ext4 partition to be clear. So I’m looking for an option to wipe the journal, like the -L option in xfs_repair. So no replay of the journal, since it stucks on it.

If it’s ext4 then you need to wait. Seems your disk is dying, so finishing the fix may take several days.

It is indeed strange, journal is tiny compared to the rest of the file system. Whatever impacts journal recovery will also likely impact any other attempt at using this file system.

Try looking into kernel logs (dmesg -w) and check whether you have I/O errors.

1 Like

This does not help you?