I want to do a shutdown -rF 3:00 to reboot my RHEL server at 3am tomorrow and force an e2fsck on each volume. But I want the e2fsck to run with the -y switch so it automatically answers yes so it will fix any problems it find automatically.

I can't find anywhere where it says if it will do this or not. does anyone know or have proof?

Or is there a better way to do what I'm trying to accomplish?

Note that some distros (eg 'buntu 9.10) have removed the -F option for shutdown
[ Check with man shutdown on your server (because you want to know which version your server will be running) ]

In which case the "new way" of forcing a fsck at the next boot is (as root):
touch /forcefsck
reboot

The default action is to fix errors automatically.

thanks tredegar. I checked and mine still has the -F option. Glad to know it will automatically fix the errors and not wait for me to press Y over and over.

Looks like I'm safe to run this then and accomplish what I want: shutdown -rF 3:00

Just another quick question, there's probably no way to run a fsck while a volume is online to fix some (but not all since the volume is mounted) errors, so when I do my shutdown -rF it could in theory take less time because some problems will already be fixed?

To be picky, the man page says time format should be hh:mm so I'd try shutdown -rF 03:00

NEVER run fsck on a mounted partition.
fsck is optimised to run as fast as it can: It will run partitions on different disks concurrently, on the same disk, sequentially. On big partitions, this can still take some time.

Be patient.

I appreciate your picky-ness, good catch.

ok, wasn't sure if i could get a head start or not.

I just looked through all the messages log files, and there haven't been any errors really at all this month. So maybe I dont even need to do the disk check? or would file system errors show up in a different log?

"there haven't been any errors really at all" or "there haven't been any errors".?

There's a big difference.

Remember that the kernel will assume the file system [ext2, ext3, ext4, reiserfs, fat32, ntfs, ... the list goes on] is clean unless the filesystem reports an error.

Errors may not be reported by an unclean filesystem. Until they become unmanageable by the filesystem. In which case the kernel may be notified "Disk full", or similar, when it clearly isn't.

If you do a search of LQ for run fsck you'll find the sort of situations that lead to this advice.

So it's a good idea to run fsck from time to time.

And I remind you that you should NEVER run fsck on a mounted partition.

But what was the problem that prompted you to start this thread about running fsck at 03:00 ?

??

Is a "head start" like a "jump start" for a car with a flat battery ?

[ And is your <SHIFT> key broken? Bad capitalisation tends to indicate .... (You probably guessed it already) ]

There aren't any errors, just status messages, in the messages logs as far back as they go.

I started this thread because we had major file system problems on this server last year, which resulted in fsck's running for almost a whole day on the server and its terabytes of data throughout. It had been years since it last fsck'd before that event, so i'd like to do them regularly now to help avoid such drastic downtime. Sounded like a good idea to me.

Where I come from, a "head start" is doing some smaller tasks before really getting into the meat of a task. Which probably wont make sense either. Say you're going to clean your house on Saturday, you could get a "head start" by going to the store Friday and buying all your cleaning supplies, saving you time Saturday. "Jump starts" are the same here as they are to you :-)

Sorry I missed my shift key starting my last sentence, I will replace my keyboard in case its faulty, LOL ;-) I have no idea what bad capitalization tends to indicate, or capitalisation as you spelled it since we're from different regions of the world. I would think it indicates I forgot to hit shift before I typed my first letter.

Thanks for the help and entertaining comments tredegar, always appreciated.

Thanks for the follow-up, hope you get it fixed ( and that it doesn't need to run for hours! )