My solaris 10 is 100% full, Any help

[gamerwalt]

Hi there, I just started with a company and i'm new to solaris. Our root is 100% full and I have no idea on what to do.

We have tried opening log and tmp files to delete thier content but we still get the same problem.

We would like to restart but I'm hoping it would not be a problem.

Any advice??

 

[KenCunningham]

I take it /var is part of the / filesystem? That's often the culprit in these cases. Check /var/adm/wtmp and /var/adm/wtmpx and shrink them with cat /dev/null > /var/wtmpx and cat /dev/null > /var/wtmp

Are there any core files? Do a find / -name core and delete as necessary. Hope this helps

Some days are diamonds, some days are rocks - make sure most are the former.

 

[bfitzmai]

Also check /var/crash/<node name>... If your system has crashed, a crash dump up to the size of your RAM will be stored there. If you are not interested in any crash analysis, you can delete all files.

 

[maximusGeek]

Do you have auditing turned on? (assuming /var is part on /) check /var/audit. audit files are very compress-able just don't compress the current audit file. You can compress and archive them according to your local log retention policy.

 

[Annihilannic]

du -kod / | sort -rn | head -20 should show you which directory is the main culprit.

Annihilannic.

 

[dandan123]

Have you tried du -skd to see if the / filesystem is really full ?

If du -skd does not show the / filesystem to be full there are two possibilities -

1. Some process has an open file which is consuming a lot of space.
2. A filesystem is mounted on a directory which has files on it and not visible due to the filesystem being mounted on it.