Drive Critical: /dev/loop0 (/var/tmp)

[Divvy]

hi friends,

Can someone please help me?
All day Im received this email from my dedicated server:
Drive Critical: /dev/loop0 (/var/tmp) is 91% full
Drive Critical: /dev/loop0 (/var/tmp) is 100% full
etc...

Some months ago I increased the /tmp partition from 512MB to 2GB, but I think that wasnt enough, because recently I am receiving the notifications again, and when that happens, my cpu load goes to 40/50, very high...

I did a:
cd /var/tmp
ls > list.txt

this is the content of list.txt file:
http://paste2.org/p/1323560

can you please suggest me what to do? Im very newbie in this things
and Im afraid to do something wrong...

Thank you

Divvy Reviewed by Divvy on 24th Mar 2011. Drive Critical: /dev/loop0 (/var/tmp) hi friends, Can someone please help me? All day Im received this email from my dedicated server: Drive Critical: /dev/loop0 (/var/tmp) is 91% full Drive Critical: /dev/loop0 (/var/tmp) is 100% full etc... Some months ago I increased the /tmp partition from 512MB to 2GB, but I think that wasnt enough, because recently I am receiving the notifications again, and when that happens, my cpu load goes to 40/50, very high... Rating: 5

[XSLTel]

rm -rf /var/tmp/*

[DLow]

I would advise maybe seting up a cron to clean it once a week =]

[Divvy]

rm -rf /var/tmp/*

What that command do?


DLow, the problem is that this happens 4 or 5 times per day...
And automatically cleans after a few minutes, but is enough to get problems with server load...

[Sp32]

The command removes all files in the folder /var/tmp

[XSLTel]

Edit : dupe

[Divvy]

is not a solution for me... the system automatically removes the files...
I only want to prevent the load thing...

cpanel support suggest me this:

It appears that something was writing to /tmp and filling it up, but it appears now to be clearing out. The df cache is still catching up at this time. It is normal for the df cache to be a bit behind on a linux server. I suspect that mysql may be behind the filling of /tmp. I'd recommend setting the tmpdir directive in your /etc/my.cnf to a larger partition. For example, something like the following:
tmpdir = /home/mysql

You should make sure that the directory is owned mysql:mysql and then restart mysql in order for this change to take effect.

What do you think?

[Lifetalk]

This command:

du -m *

Will show you the filesize of each individual file in that directory. I'm assuming the .myd files towards the end are the largest ones, since session files are barely a few KBs.

Depending on what you find out from the above command, you can create a cron script to delete those files, say, every hour. Or, if you can figure out what program is generating them, you could change the location they are generated in. And if that isn't possible, then you could prevent them from being generated at all.

OR, if all else fails, increase the size of your tmp partition again.

[Divvy]

thank you for your reply lifetalk.
Btw, do you think that increasing the partition size will solve my problem?
I think that the cron job to delete files every hour will not work, because I notice that the loads increase when files is copying to the partition folder.

I think that the best way is to find what script is causing this...
I have checked my whm logs and found this:
http://i.imgur.com/Pt33f.png
http://i.imgur.com/uDwM8.png
http://i.imgur.com/xE9Qo.png

As you can see, the users sapo and tugapira is the main problem, right?
Both users using wordpress with the same exactly plugins. Could be a plugin that is causing this? If yes, what is the better way to find the guilty one? Or not possible?

Please help, I'm desperate...

[Gavo]

Follow what cpanel support told your to do, or increase /tmp size

http://forums.cpanel.net/f5/how-resize-tmp-86345.html