2.2.14 - grow_inodes: inode-max limit reached - how to find/fix the inode leak?

From: Peter Gervai (grin@tolna.net)
Date: Tue Apr 25 2000 - 10:26:27 EST


Kernel is 2.2.14+openwall_secure patch. In 5 hours uptime I have the following:

root@melanie:~# cat /proc/sys/fs/inode-*
31950 5
31950 5 0 0 0 0 0

(65535 was set to prevent the need to reboot. *sigh*)

root@melanie:/dev# cat /proc/sys/fs/file*
1529 215 8192

The system allocates 30 new inodes per 20 seconds on average.

Reaching the limit results the nice messages in the logfile:

Apr 25 09:30:18 melanie mon[768]: Could not fork
Apr 25 09:30:19 melanie kernel: grow_inodes: inode-max limit reached
Apr 25 09:30:31 melanie last message repeated 2 times
Apr 25 09:30:31 melanie kernel: socket: no more sockets

Which can be fixed by the three finger salute. If someone's around.
(Well actually not, since init 6 cannot be forked either.

Someone's eating up my inodes, but I see only ~4000 open FD's around
(whether this matter in this case or not), and I cannot spot any
runaway processes. The machine does not have lusers so it isn't
someone playing rude games. Everything

Questions: Can I spot who's eating up inodes? (lsof doesn't give any
hint, nor did I expect it.)

What can I do apart from rebooting the server every day?


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

This archive was generated by hypermail 2b29 : Sun Apr 30 2000 - 21:00:09 EST