2

Most docs on disabling/creating the ext4 journal for root involve a reboot. If we:

  • turn off the journal via tune2fs
  • fsck the filesystem
  • turn the journal back on via tune2fs
  • fsck the filesystem
  • then reboot

Will that effectively re-create the journal? In this specific case we have a corrupted journal we wish to re-initialize. If we can eliminate the intervening reboot then we can better automate this process.

I don't actually see why a reboot is required at any point?

Thanks.

user152471
  • 21
  • 2

1 Answers1

2

Your point would be excellent if it were not for the logical impossibility in steps two and four. You cannot meaningfully fsck a mounted filesystem; they are made self-inconsistent by the very act of being mounted. You can't unmount the partition to fsck it, because it's the root partition. The fsck requirement, I suspect, is why all instructions involve a reboot or two.

MadHatter
  • 79,770
  • 20
  • 184
  • 232
  • "mount -o remount" in single user mode? – user152471 Jan 08 '13 at 10:07
  • `mount -o remount,ro` would probably do it, but it'll essentially stop the machine from working. I'd assumed you wanted to avoid a reboot because of service interruption; if you don't mind service interruption, why not reboot and do it cleanly? – MadHatter Jan 08 '13 at 10:10
  • A service interruption is fine, the purpose is to better automate the process. – user152471 Jan 08 '13 at 10:18
  • OK, then modify your procedure above; looks like we might be iterating towards a solution! – MadHatter Jan 08 '13 at 10:19