I have two servers (a file store and an email server) where winbind intermittantly stops running. I have kept a log of these times, and examined them for a pattern. I beleive I have traced the fault on the email server to the time the kerberos ticket is renewed. I have adjusted the time as follows ticket_lifetime = 36000 renew_lifetime = 40000 This has slightly improved the reliability as it now renews two out of three times successfully. But on the third time winbind dies, if left samba eventually crashes. I have put a script in place which checks the winbind service and restarts it if it is not running. I would appreciate any help or guidance in this matter. Ian
Could you possibly * test whether this is still the case with 3.0.23rc2 * provide us logfiles, gdb backtraces ?
* test whether this is still the case with 3.0.23rc2 * provide us logfiles, gdb backtraces ? I will download and install 3.0.23rc2 hopefully today. I am fairly new to linux, please could you tell me which logfiles you require, where do I send them too. How do I get a gdb backtrace? Ian
I will not be upgrading to samba 3.0.23rc2 as this release says it is not for production servers.
We're not asking you to update to 3.0.23RC code on production servers, but to test it on non-production ones. That way you're helping us get it ready for production use. Thanks, Jeremy.
The tgt != tgs ticket lifetime has been adressed in bug #4009 and will be available in 3.0.25 (or your vendor picks up the patch earlier). *** This bug has been marked as a duplicate of 4009 ***