OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
Re: 8000 active, rndc flush cleared it

From: Wietse Venema (wietseporcupine.org)
Date: Wed Aug 15 2007 - 09:23:54 CDT


Len Conrad:
> fbsd 6.2
> postfix 2.5-20070731
> BIND 9.4.1-P1
>
> 8000+ msgs in active q. Only a handful in deferred. Situation stayed
> like that for 1 to 2 hours. The active msgs were both inbound to our
> next-hop of content-scanner boxes, and outbound as SAV
> probes. Another MX sharing the load was delivering successfully to
> the c-s boxes.
>
> maillog showed no panics or errors
>
> dmesg -a showed nothing
>
> /var/log/messages showed nothing
>
> top had the machine at 95% idle, lots RAM free or inactive.
>
> only 30 smtpd processes
>
> rndc status showed less than 100 active recursive queries.
>
> "rndc flush" allowed postfix to drain the active queue. Still
> working 13 hours later.
>
>
> Any suggestions what was blocking postfix?

Exponential backoff to maximal_backoff_time.

        Wietse