[Bug 1398] DHCP server stops unexpectedly after less then 24 hours

[drift@skolelinux.org]

http://bugs.skolelinux.org/show_bug.cgi?id=1398


John Sigurd Skogtvedt <jss@bzz.no> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jss@bzz.no




--- Comment #10 from John Sigurd Skogtvedt <jss@bzz.no>  2009-12-20 10:33:43 ---
(In reply to comment #9)
> I've been able to reproduce this, and the problem is SIGPIPE.  dhcpd dies with
> SIGPIPE when trying to talk to the LDAP server.  I believe this only happen if
> no LDAP request have been submitted for a long time, leading to the server
> closing the connection.  The dhcpd server should then reconnenct, but instead
> if dies with SIGPIPE.
> 

A possible fix could be to write a LD_PRELOAD'ed .so which installs a SIGPIPE
handler which either restarts dhcpd or sends a signal (e.g. SIGHUP) which
causes it to reload the configuration file and thus open a new LDAP connection.
Haven't had the opportunity to look into the details to check if this would
work. It is slightly hacky, but would not require patching dhcpd itself.

Another thought: nagios-plugins has a check_dhcp plugin which is used to test a
DHCP server. Perhaps it could be used to send a unicast DHCP request every five
minutes from a cronjob.


-- 
Configure bugmail: http://bugs.skolelinux.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
You are watching all bug changes.