Arch Linux

Description:
nscd segfaults | possibly related to s2ram?

Additional info:
* package version(s)
core/glibc 2.11.1-2

* config and/or log files etc.
$ grep -i nscd everything.log
...
Apr 17 12:25:21 cinnabar kernel: nscd[18434]: segfault at 10 ip b776f238 sp bfbbd2a0 error 6 in nscd[b776b000+1c000]
Apr 17 12:34:37 cinnabar kernel: nscd[18475]: segfault at a ip b76f3cdb sp aef759dc error 4 in libc-2.11.1.so[b7680000+141000]
The entire output is attached

Using default /etc/nscd.conf
Using modified /etc/nsswitch.conf, (for avahi mdns) attached

$ sudo rm -f /var/db/nscd/* /var/run/nscd/*; sudo ls -lah /var/db/nscd /var/run/nscd
$ sudo -u root bash -c "/usr/sbin/nscd --debug" |& tee nscd.debug.option
Suspend2ram for about two hours, resume, crash. The file is attached

Steps to reproduce:
start nscd & wait. It seems there is a relation between the nscd segfaults and suspend2ram (via uswsusp, if it makes a difference) - shortly after resume I've seen nscd segfault. Since not all the nscd segfault times logged by the kernel correlate 1:1 with when I've resumed from sleep, this theory needs more testing. This'll take some time though: the segfaults seem only to occur if the computer been suspended long enough, at least "overnight", possibly less.

Here are some possibly related bugs
. opened by me, dealing with privoxy
https://sourceforge.net/tracker/index.php?func=detail&aid=2962037&group_id=11118&atid=361118#
. chromium issue
http://code.google.com/p/chromium/issues/detail?id=11380

Idea: Could nscd be adversely affected by clock changes? I have a dispatcher script "/etc/NetworkManager/dispatcher.d/openntpd-dispatch" that changes the clock ("/usr/sbin/ntpd -s") on NetworkManager up/down events. I'll test this