/usr 100% full

I'm not the person to be giving advice, but when I had that very issue, it turns out I had a bunch of core dumps (core.12345, core.11111, etc) that were taking up space. Search the forums here to learn more on that. Core dumps are key for diagnosing what was wrong but mine were old, dating back to when I had memory issues on the server. Deleting those old core dumps did part of the trick. The other part was finding and dealing with the cpbackup folders, which I had turned off, but apparently some were there anyway. Those backup folders can take up some serious space and may be there long after you've turned off that setting in Tweak Settings - if you did so.

There's also a post by Chirpy talking about moving your
/usr/local/apache/domlogs
and linking back to them.
I didn't need to do that log moving/syncing because I freed up enough space doing the first two-- finding the obsolete core dumps and cpbackups and deleting them.

 

Use
du --max-depth=1
in the /usr directory to find out which subdirectory is using the most space, then do that from that directory and so on to find out where the space is being used.

 

The /usr/lib directory is used for program running files. There's probably not a lot there to remove. What about the log files?
/usr/local/apache/logs
/usr/local/apache/domlogs
/usr/local/cpanel/logs

That tends to be the cause of low space there a lot.

 

If you are running mailman service, take a look in

/usr/local/cpanel/3rdparty/mailman/logs

and

/usr/local/cpanel/3rdparty/mailman/archives

 

how much space have you allocated for /usr ?

 

What OS are you running on your new server? You didn't say. For example, Centos 4.4 uses about 3Gb in /usr, so 9Gb seems like ample to me.

Another command you could run is:

Code:

find /usr -size +4000 -mtime -10 | xargs ls -ld

That will run ls -l on any large files (> 2mb) created in the last 10 days which may help you track down a runaway log or other file.

The next step might be something like this:

Code:

cd /usr
du -sk *| sort -nr

which will display a list of directory sizes. For each bigger sized directory, cd into that directory and rerun the "du -sk *" to see what the bigger culprits are. However - BE WARNED - if you don't know what the files are, you'd be insane to delete them, you could end up with a badly broken system. There's no other way to say this - administering a Linux system is a skilled process, and you're going to need occasional help.

 

Just another thought - perhaps whoever you got the server from couold be persuaded to fix it. After all, they provided it to you in a broken state, and it should have been provided in a working state.

 

I know no one can answer fast enough when someone is looking for a solution but could you post a follow up as to the trouble your data center found? It would be greatly appreciated and perhaps give us some insight for future situations like this while helping blindly.