On Mon, Oct 19, 2009 at 10:48 AM, Brian McCann <bjmcc...@gmail.com> wrote:
> Hi all. I'm having a problem with relayd under OpenBSD 4.5. I've got
> it configured to load balance web traffic (almost exactly the
> configuration in example 1 at https://calomel.org/relayd.html ). It
> was working great until I increased the load...then odd things started
> happening. relayd would exit (I have it running under daemontools,so
> it would start back up), with the following in the log:
>
> --snip--
>
> 2009-10-19 10:35:17.647251500 kill_tables: deleted 0 tables
> 2009-10-19 10:35:17.647410500 host check engine exiting
> 2009-10-19 10:35:17.650175500 relay iisproxy, session 281 (3 active),
> 0, 24.149.127.253 -> 192.168.9.3:80, shutdown
> 2009-10-19 10:35:17.650473500 flush_rulesets: flushed rules
> 2009-10-19 10:35:17.651681500 pf update engine exiting
> 2009-10-19 10:35:17.651728500 relay iisproxy, session 277 (2 active),
> 0, 24.149.127.253 -> 192.168.9.4:80, shutdown
> 2009-10-19 10:35:17.653318500 relay iisproxy, session 286 (1 active),
> 0, 199.191.107.2 -> 192.168.9.2:80, shutdown
> 2009-10-19 10:35:17.664604500 socket relay engine exiting
> 2009-10-19 10:35:17.664654500 socket relay engine exiting
> 2009-10-19 10:35:17.664698500 socket relay engine exiting
> 2009-10-19 10:35:17.664742500 socket relay engine exiting
> 2009-10-19 10:35:17.664786500 socket relay engine exiting
> 2009-10-19 10:35:17.664829500 socket relay engine exiting
> 2009-10-19 10:35:17.664873500 socket relay engine exiting
> 2009-10-19 10:35:17.664917500 socket relay engine exiting
> 2009-10-19 10:35:17.664961500 socket relay engine exiting
> 2009-10-19 10:35:17.665072500 socket relay engine exiting
> 2009-10-19 10:35:17.665118500 socket relay engine exiting
> 2009-10-19 10:35:17.665162500 socket relay engine exiting
> 2009-10-19 10:35:17.665206500 socket relay engine exiting
> 2009-10-19 10:35:17.665251500 socket relay engine exiting
> 2009-10-19 10:35:17.665297500 socket relay engine exiting
> 2009-10-19 10:35:17.665342500 socket relay engine exiting
> 2009-10-19 10:35:17.665387500 socket relay engine exiting
> 2009-10-19 10:35:17.665553500 socket relay engine exiting
> 2009-10-19 10:35:17.665599500 socket relay engine exiting
> 2009-10-19 10:35:17.665643500 socket relay engine exiting
> 2009-10-19 10:35:17.665695500 socket relay engine exiting
> 2009-10-19 10:35:17.665738500 socket relay engine exiting
> 2009-10-19 10:35:17.665781500 socket relay engine exiting
> 2009-10-19 10:35:17.665819500 socket relay engine exiting
> 2009-10-19 10:35:17.689529500 startup
>
> --/snip--
>
> ...then it would start back up. To make things more interesting, the
> IP stack would appear to go insane. I was getting the following
> logged in relayd:
> --snip--
> relay_connect: session 50064: forward failed: No route to host
> --/snip--
>
> I could also no longer ping any local addresses, nor the local
> interface's IP address itself...both would fail with "no route to
> host". I figured it may have been a bad NIC, but I switched over carp
> to my standby box, and the same thing immediately happened there. I
> also applied some of the performance tweaks at
> https://calomel.org/network_performance.html, but that didn't help
> any.
>
> Nothing gets sent to dmesg or /var/log/messages when this happens. I
> figure I've got to be missing some sysctl or something to make some
> buffer bigger, but I can't figure out what.
>
> Anyone have any clues here?
> Thanks!
> --Brian
>
For those following along in the archives...the problem was that I was
running pf, and the state table for pf wasn't big enough. Once I
increased that, it worked fine.
--Brian
--
_-=-_-=-_-=-_-=-_-=-_-=-_-=-_-=-_-=-_-=-_-=-_
Brian McCann
"I don't have to take this abuse from you -- I've got hundreds of
people waiting to abuse me."
-- Bill Murray, "Ghostbusters"
