[Bug 1332203] Re: autofs5-ldap regularly crashes after long periods of time

[Paul Kilgo]

This problem has continued over the past few months. We slogged through
the strace output a few times and just today found this:

[pid   935] sendto(11, "<27>Sep 22 13:06:30 automount[92"..., 90, MSG_NOSIGNAL, NULL, 0) = -1 ECONNREFUSED (Connection refused)
[pid   935] close(11)                   = 0
[pid   935] socket(PF_FILE, SOCK_DGRAM|SOCK_CLOEXEC, 0) = 4
[pid   935] connect(4, {sa_family=AF_FILE, path="/dev/log"}, 110) = -1 ECONNREFUSED (Connection refused)
[pid   935] close(4)                    = 0
[pid   935] ioctl(3, 0xc0189379, 0x7f22fecabb20) = 0
[pid   935] ioctl(3, 0xc0189375, 0x7f22fecabb20) = 0
[pid   935] sched_yield()               = 0
[pid   935] umount("/home", 0)          = -1 EBUSY (Device or resource busy)
[pid   935] nanosleep({0, 200000000}, NULL) = 0
[pid   935] umount("/home", 0)          = -1 EBUSY (Device or resource busy)
[pid   935] nanosleep({0, 200000000}, NULL) = 0

Our syslog sends some events to another loghost and the top call to
sendto(), which appears to fail, and automount tries to go down
gracefully. Checking the .crash file again, the backtrace is identical.
The backtrace does seem to support this story since there is a call to
vsyslog() inside of log_error().

So this bug may be filed under the wrong package, depending on whose
responsibility it is to handle this kind of error. Reading the man page
for vsyslog(), I don't see any way the caller can catch an error so this
could in fact be a problem with it.

We are using rsyslog 5.8.6-1ubuntu8.9 currently.

-- 
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to autofs5 in Ubuntu.
https://bugs.launchpad.net/bugs/1332203

Title:
  autofs5-ldap regularly crashes after long periods of time

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/autofs5/+bug/1332203/+subscriptions