Hi ,

I have a production server with Centos 5.2 running. Suddenly my / partition became 100% full.
Is there anyway to remove file under / which doesnot affect the production server. I have many application running under /home . Increasing /home partition would increase / parition too?



Filesystem Size Used Avail Use% Mounted on

/dev/mapper/VolGroup00-LogVol00
3.9G 3.7G 0 100% /

/dev/mapper/VolGroup00-LogVol04
52G 33G 17G 67% /home

/dev/mapper/VolGroup00-LogVol02
2.0G 344M 1.6G 19% /var

/dev/mapper/VolGroup00-LogVol03
2.0G 68M 1.8G 4% /tmp

/dev/sda1 190M 38M 143M 21% /boot

tmpfs 1006M 0 1006M 0% /dev/shm


Any help would be highly appreciated
Thanks

The Good News is that your /home, /var and /tmp partitions are all fine - you have plenty of space there.

Now you need to find out who the space hogs are in "/". What you can do about it depends entirely on what you find. It might be easy to free the space ... or it might be relatively painful. You simply have to look.

Run "du -sk" from the root "/", then look one directory down (in the root "/" partition only), etc. You probably won't have to look to far...

Good luck .. PSM

your / is only 3.9 gig in size ( that is one BIG "small" problem ) that partition is WAY to small.
your /home is 52 gig -- that is WAY to big.

repartition the disk so that / is AT LEAST 9 gig.

also a /tmp is not needed leave it on /
nor is /var you can also leave that on /

/home should also be smaller say 5 to 10 gig -- and not 52 gig

John VV -

I agree that the disk partitions could be allocated better.

But the OP said "production server". As such, I suspect the risks in repartitioning the drive probably outweigh the benefits.

But in any case, I think it makes sense to figure out WHY the root partition "suddenly became 100% full". BEFORE doing anything else.

Who knows - maybe somebody was trying to do a backup, and created a 2GB tar file in /dev? For this example, all he'd have to do is "rm" the file, and the problem would be resolved.

In any case, I'd definitely make an effort to determine the "space hog(s)" first. It's simple, and it's time well invested.

IMHO .. PSM

well my CentOS 5.3 / is 4.1 gig
yes i do have Gnome and KDE installed there

the op's /var might be the logs , if cron is not cleaning the old ones then/var/log/* can get BIG
the same thing for /tmp a cron script is not cleaning out old things

What could be the reason for not cleaning out old logs by cron?

Backup data.
Repartition / .

I assume your filesystems are ext3 so by default 5% is reserved so at 100% its >95% of actual data. As suggested you need to find out why / is full.

BTW the server is using LVM so it is much easier to resize a logical volume then it would be to resize a regular partition. As always backup any and all essential data.

You might of installed something to /usr which then filled up /.

having it turned off so that the 30 day delete is not ran .
or 30 days is to long and it needs to be set to 7 days .

Specifically, check logrotate jobs/settings

...