/tmp running out of inodes
Tankko
tankko at gmail.com
Thu Dec 18 17:25:32 UTC 2008
For the second time in 2 weeks my server has gone down due to running
out of inodes in /tmp. The first time it did this /tmp was filled
with files, so that made sense. I delete all the files from /tmp and
all was well for a few weeks. Just this morning it did it again (keep
in mind that this sever is currently 700 miles away) and had to be
rebooted.
But when I looked at /tmp, it was empty except for a couple of files,
but running df shows the follow:
Filesystem Size Used Avail Capacity iused ifree %iused Mounted on
/dev/ad4s1a 496M 59M 398M 13% 1535 64255 2% /
devfs 1.0K 1.0K 0B 100% 0 0 100% /dev
/dev/ad4s1e 496M 69M 387M 15% 65790 0 100% /tmp
/dev/ad4s1f 221G 5.2G 198G 3% 337742 29596848 1% /usr
/dev/ad4s1d 1.9G 757M 1.0G 42% 1632 280990 1% /var
/dev/md0 19M 12K 18M 0% 9 2805 0% /tmp
/dev/ad5s1c 289G 49G 217G 18% 27779 39115643 0% /bigdrive
65790 used inodes and 0 free on /tmp
I ran fsck and it shows:
** /dev/ad4s1e (NO WRITE)
** Last Mounted on /tmp
** Phase 1 - Check Blocks and Sizes
** Phase 2 - Check Pathnames
** Phase 3 - Check Connectivity
** Phase 4 - Check Reference Counts
** Phase 5 - Check Cyl groups
65790 files, 35374 used, 218441 free (9 frags, 27304 blocks, 0.0% fragmentation)
So where are these mysterious 65790 files? Is there a way to clean
this up? And more importantly is there a way to do this remotely
given that I am 700 miles from the server right now?
Every time the server goes down due to this problem, it won't come up
again without someone physically running fsck at boot.
On a side question...what the the best policy for deleting files from
/tmp? Seems like a lot of apps are happy to leave files in /tmp. Is
clean up commonly done as a cron job? What about files like
mysql.sock= which are important. I can't just blindly remove
everything in /tmp each night.
Thanks for any help.
Tankko
More information about the freebsd-questions
mailing list