Red Hat Bugzilla – Bug 114969
tmpwatch does not clean up socket files in /tmp
Last modified: 2007-11-30 17:06:53 EST
Description of problem:
tmpwatch does not remove socket files, in particular, files in
/tmp/orbit-*/linc* left by gnome were allowed to grow.
Version-Release number of selected component (if applicable):
tmpwatch-2.8.4-5 (RHEL3, RH9) down to tmpwatch-2.8.3-1 (RH7.3)
Steps to Reproduce:
1. login via gdm
2. use the system for some period of time
socket files accumulated in /tmp/orbit-*/*
clean up after a while.
Adding --all in /etc/cron.daily/tmpwatch would be a solution, i.e.
/usr/sbin/tmpwatch --all 240 /tmp
/usr/sbin/tmpwatch --all 720 /var/tmp
This is too dangerous.
It could for example remove the sockets used for X11
(atime on the socket is changed only on new connections,
not only on X protocol traffic).
*** This bug has been marked as a duplicate of 66148 ***
Right, a safer version that I have been using is the following:
# cleanup accumulated socket files in /tmp/orbit-* first
/usr/sbin/tmpwatch --fuser --all 240 /tmp/orbit-*
/usr/sbin/tmpwatch 240 /tmp
/usr/sbin/tmpwatch 720 /var/tmp
The rationale from bug 66148 is summarized in comment 1 above.
FWIW, I've seen filesystems run out of inodes due to this problem.
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.