When multipathd tries to fork a callout, if it fails, it doesn't close the pipe fds necessary to talk to the callout. Also, when multipathd fails to execute a callout, it doesn't print any useful error messages.
Is RH planning to fix this for RHEL5.3? We need to determine how much longer our test bed/test resources will be consumed with this release and whether we need to regression test another RHEL5.3 RC?
Laurie
The fix for this is committed. The plan is to release this as a z-stream release. Although I noticed this bug looking at 478643, as far as I know, nobody has ever hit it. The error reporting is very helpful to tell if people are running into 478643, and need to increase their max_fds, but as long as you follow the workaround in that bugzilla, you should be fine.
An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.
http://rhn.redhat.com/errata/RHEA-2009-1377.html