OK, I'm getting some of the "Poll failed to start" messages too. I'll have to poke around in code and see what that really means.
Aug 28 11:45:25 slbd[69105]: Using r_refresh of 5000 milliseconds Aug 28 11:45:25 slbd[69105]: Using configuration file /var/etc/slbd.conf Aug 28 11:45:25 slbd[69105]: VIP 10.0.2.66:80 configured as "10.0.2.66" Aug 28 11:45:25 slbd[69105]: VIP 10.0.2.66:80 sitedown at 10.0.2.66:80 Aug 28 11:45:25 slbd[69105]: VIP 10.0.2.66:80 added real service 192.168.10.2:80 Aug 28 11:45:25 slbd[69105]: VIP 10.0.2.66:80 added real service 192.168.10.3:80 Aug 28 12:31:59 slbd[69105]: Poll failed to start TCP to 192.168.10.3:80 Aug 28 12:32:04 slbd[69105]: Poll failed to start TCP to 192.168.10.3:80 Aug 28 12:32:04 slbd[69105]: Poll failed TCP for 192.168.10.3:80 On 8/28/05, Matthew Lenz <[EMAIL PROTECTED]> wrote: > ok I take some of that back.. I'm getting the poll failures on the primary > as well. > > Side/unrelated note. Does carp log any where when it switches an interface > from master to backup? > > ----- Original Message ----- > From: "Matthew Lenz" <[EMAIL PROTECTED]> > To: "Scott Ullrich" <[EMAIL PROTECTED]> > Cc: "pfsense" <[email protected]> > Sent: Saturday, August 27, 2005 9:35 PM > Subject: Re: [pfSense-discussion] load balancing failures > > > > (sorry for the top reply) > > > > It also seems to be messing with carp failover as well. Its really > > strange.. everything seems to work fine on my primary (no preempt enabled > > currently) but I if I try to run the EXACT same config (primary has a > > advskew of 0 and secondary has a advskew of 1, but that should only matter > > if preempt is enabled) on my secondary i get the lb errors and soon after > > i get carp failures. Its hard to know really whats causing it though > > because I upgraded to 80.2 yesterday and started using slbd at the same > > time. I need to get down to the datacenter and see if disabling some of > > the extra onboard hardware helps matters. Does freebsd do smp very well? > > this is a hyperthreaded intel cpu in these machines and it looks like its > > enabling smp in the syslog. > > > > ----- Original Message ----- > > From: "Scott Ullrich" <[EMAIL PROTECTED]> > > To: "Matthew Lenz" <[EMAIL PROTECTED]> > > Cc: "pfsense" <[email protected]> > > Sent: Saturday, August 27, 2005 9:13 PM > > Subject: Re: [pfSense-discussion] load balancing failures > > > > > > On 8/27/05, Matthew Lenz <[EMAIL PROTECTED]> wrote: > >> I'm getting a lot of: > >> > >> Aug 27 20:42:21 slbd[684]: Poll failed to start TCP to 192.168.1.2:443 > >> Aug 27 20:42:21 slbd[684]: Poll failed TCP for 192.168.1.2:443 > >> Aug 27 20:42:21 slbd[684]: Poll failed to start TCP to 192.168.1.3:443 > >> Aug 27 20:42:21 slbd[684]: Poll failed to start TCP to 192.168.1.3:80 > >> Aug 27 20:42:21 slbd[684]: Poll failed TCP for 192.168.1.2:80 > >> Aug 27 20:42:21 slbd[684]: Poll failed to start TCP to 192.168.1.3:80 > >> Aug 27 20:42:26 slbd[684]: Switching to sitedown for VIP 12.34.56.78:443 > >> Aug 27 20:42:26 slbd[684]: Switching to sitedown for VIP 12.43.56.78:80 > >> > >> seemingly at random. These servers are up and running just fine so I'm > >> not > >> sure what the problem might be (yes i can even hit them through the VIP > >> over > >> the internet and load balances just fine). How do I go abou debugging > >> this > >> to give you guys a better idea of what might be wrong? Slbd's logging is > >> kind of unhelpful as well. It logs every 5 seconds that the poll fails > >> but > >> then doesn't log when its back up. It seems it would be better if it > >> logged > >> when it went down and when it came back up. > > > > > > This appears more of a "I cannot even open a socket error" other than > > "I cannot connect to the target". > > > > SLBD apparently needs some more work. We had to port a lot of changes > > from it being against a earlier version of pf. > > > > Scott > > >
