View previous topic :: View next topic |
Author |
Message |
jasenux n00b
Joined: 22 Jul 2007 Posts: 12
|
Posted: Thu Mar 08, 2012 7:50 am Post subject: help needed - keepalive + haproxy |
|
|
Hi all,
Help needed.
I have 5 different circuits, and on each circuit, I have a pair of load balancer (running haproxy 1.4.8 or 1.4.19 + keepalived 1.1.20 or 1.2.2-r3 on Gentoo). 4 pairs out of these 5 work perfectly. Only the fifth pair is not working properly. They all share the same configuration - except of course, the IPs are different.
The fifth pair will work for a bit (30 min to 4 hours) then die. No issues with the other 4 pairs. With the fifth pair, if I just run haproxy, it is fine. They are all VMware guests. The only difference is how they connect to the Internet. The good 4 pairs connect to the Internet via DIA; while the fifth one is behind a IP-IP GRE tunnel (hence MTU is 1476 and the MSS is 1436). When it is not working, I don't even see packets coming to it - I do see packets passing the firewall and heading towards the load balancer. If I restart keepalived again, I will start seeing packets again....
keepalived.conf:
vrrp_script check_haproxy {
script "killall -0 haproxy"
interval 2
weight 2
}
vrrp_instance VI_174 {
interface eth0
state MASTER
virtual_router_id 174
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 162163
}
virtual_ipaddress {
a.b.c.174
}
track_script {
check_haproxy
}
} |
|
Back to top |
|
|
NewBlackDak Guru
Joined: 02 Nov 2003 Posts: 512 Location: Utah County, UT
|
Posted: Tue Oct 16, 2012 2:26 pm Post subject: Re: help needed - keepalive + haproxy |
|
|
jasenux wrote: | Hi all,
Help needed.
I have 5 different circuits, and on each circuit, I have a pair of load balancer (running haproxy 1.4.8 or 1.4.19 + keepalived 1.1.20 or 1.2.2-r3 on Gentoo). 4 pairs out of these 5 work perfectly. Only the fifth pair is not working properly. They all share the same configuration - except of course, the IPs are different.
The fifth pair will work for a bit (30 min to 4 hours) then die. No issues with the other 4 pairs. With the fifth pair, if I just run haproxy, it is fine. They are all VMware guests. The only difference is how they connect to the Internet. The good 4 pairs connect to the Internet via DIA; while the fifth one is behind a IP-IP GRE tunnel (hence MTU is 1476 and the MSS is 1436). When it is not working, I don't even see packets coming to it - I do see packets passing the firewall and heading towards the load balancer. If I restart keepalived again, I will start seeing packets again....
keepalived.conf:
vrrp_script check_haproxy {
script "killall -0 haproxy"
interval 2
weight 2
}
vrrp_instance VI_174 {
interface eth0
state MASTER
virtual_router_id 174
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass 162163
}
virtual_ipaddress {
a.b.c.174
}
track_script {
check_haproxy
}
} |
Zombie post I know, but we're seeing the same thing with a similar setup.
2 identical vms:
gentoo-sources-3.1.6
keepalived-1.2.2-r3
haproxy-1.4.18-r1
1 vCPU
2 GB RAM
Haproxy just runs forever on both machines. We introduced keepalive after a storage mishap. The site was down for nearly 15 minutes while we rebooted the haproxy box, and waited for fsck. Keepalived runs anywhere from 4 days to around 6 weeks, and dies without any warning. Restarting keepalived brings the site back online immediately. We have resorted to rebooting them on regular intervals. It's not ideal, but we need 99.999% up time for this service, and it's our only workaround at this point. _________________ Gentoo systems.
X2 4200+@2.6 - Athy
X2 3600+ - Myth
UltraSparc5 440 - sparcy |
|
Back to top |
|
|
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|