Skip fsck on reboot

[kenny chia]

Hi
I have performed reboot of linux servers (Red Hat 6.X) that have been up for one year and encountered cases of fsck failing. After fsck failed, all I could do is to go into maintenance mode or reboot. Worse of all, fsck further corrupted the filesystems

Since the server has been running without fail or problems, will it be better to skip fsck on reboot?

[Vitaly Karasik_1]

I suggest you to go to maintanance mode and run fsck manually.

I don't think skipping fsck is a good idea.

Vitaly.

[Zeev Schultz]

Have you probably noticed what fsck was
complaining about?Have you also noticed what
partition was it complaining about?
The unwritten rule is to boot the server into
single-user mode with only limited number of filesystems mounted and run fscks as much as needed.

[I_M]

Hi,

Linux ext2 filesystem is being fscked either
* Exceed maximum mount count.
* Exceed interval-between-checked.

So you may modify these valuses using /sbin/tune2fs command.

Check tune2fs online manual for more info.

Good luck

[Balaji N]

hi

if fsck tries to scan, it means that either there is a problem with the filesystem or it has not updated the filesystem after fixing the problem previous table.

suggest to boot from a cd and manually run fsck.

also, RH6.x belongs to stone age. consider upgrading. and use a journalling filesystem like ext3 to mimise hassles.

-balaji

[kenny chia]

I have tried to do fsck manually.. It "fixed" the filesystem but it ended up corrupting the filesystem instead. For example, it reassigned files to another name or the files simply became inaccessible.

I'm stuck with too many stone age Red Hat 6.X servers that I cannot upgrade..

The only solution to my problem so far is
1. Identify the corrupted partition.
2. Backup the entire partition.
3. Delete all files in the partition. This will remove the corrupted files
4. Restore back the files in the partition

[Sergejs Svitnevs]

In this situation i would remove the harddrive of an other machine which
is running the same Redhat version and put it in the old linux server. I would use this harddrive to boot linux and repair the file system on
the original harddrive (just copy the files off the other machines harddrive).

Regards,
Sergejs

[Zeev Schultz]

Kenny,
are whose system or user files?most probably
fsck tried to repair some lost/unassigned inodes-files after a bad/unclean shutdown.
There are also ways to repair the files found
in lost+found depends on damage.
Zeev

[Steven E. Protter]

Unix Systems really,really,REALLY need to run fsck to maintain logical volume(HP-UX) and filesystem health.

Linux is no exception.

If fsck if finding problems and messign up filesystems then maybe its the problem and you need to get new binaries, or do an OS update.

Just because the system is running right now doesn't mean it will forever. You need to track down why this is happening and correct it.

I think this could be a mirroring problem, or your disk is giving you troble, but I can not say for certain.

If its redhat, see if there is an update for fsck and other OS disk utilities.

SEP

[Lior Mishkovsky]

Hi,

Running on a non a fsck'd system is like running into a hole in the groud - eventually you'll have to face the problem, better do it now.
My advice is to backup the faulty system,check if there are also IO error's in the syslog that shows physicall problems on the disk.
If so replace the disk and reinstall the os - later restore the data.
In case of only SW/OS problem
reinstall the OS with repartioning the disk.
If it's a production 24*7 system - create an alternate server beforehand.

And remember - backup first, think later...

Best Regards.
Lior./