Had two crashes today so far and I've managed to pull some more info from the logs and reports.
We use the squid Kerberos authentication plugin to authenticate the user and their password, then we use the wbinfo_group.pl script to check the users membership of an AD group. It appears that the checking for a group stage is where winbind crashes. Interestingly the Kerberos authentication stage carries on authenticating users properly even after the crash. Users that have had their authentication cached by squid can carry on accessing the internet as normal for a short time. Any new authentication attempts fail. The SID in each initial segfault report was the SID for the "Domain Users" group. The SIDs in the following segfaults are a mix of random groups and users, different after each crash. The last successful user to be looked up before each crash was also different. After the first segfault I had to force a server restart as it became totally unresponsive. The second time I managed to kill the Winbind service and restart it. No other faults so far but it does seem to be a quiet period at the moment, so the server isn't under a lot of load. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1100787 Title: Samba 3.6.3 segfault or panic. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/samba/+bug/1100787/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs