LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

[lvs-users] Checkinterval issue.

To: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Subject: [lvs-users] Checkinterval issue.
From: Piotr Gbyliczek <pg@xxxxxxxxxxxxxx>
Date: Thu, 26 Jun 2008 16:58:06 +0100
Hi

First, greetings to all on this list, I've just signed in.

Second, the issue.

I have load balancing solution, using ldirectord to balance http 
requests between farm of servers. Almost everything is working fine. 
Servers death is detected, failover works, all looks nice. Problem I 
have is that load balancer is checking apache status page only like 
every 4 minutes, while checkinterval is set to 10 seconds as far as I 
can understand this setting. This means that apache on one real servers 
can be stopped, but request are still getting to it, as it is not 
detected as faulty. Any ideas why it is like that ?? Any places where to 
search ?? I was searching on Google, but nothing useful came out - all 
issues were because this parameter was not set at all, so it was 
defaulting to 30 seconds.

OS is Centos 5.1 64 bit, heartbeat and ldirector taken from default 
repositories.

Here is ha.cf file :

logfacility     local0
auto_failback on
debug 1
keepalive 2
deadtime 10
warntime 10
initdead 10
mcast eth0 225.0.0.1 694 1 0
mcast eth1 225.0.0.1 694 1 0
node    lba1
node    lba2
respawn hacluster /usr/lib64/heartbeat/ipfail
apiauth ipfail gid=haclient uid=hacluster


and part of ldirectord.cf :

# Global Directives
checktimeout=5
checkinterval=10
fallback=127.0.0.1:80
autoreload=yes
logfile="/var/log/ldirectord.log"
#logfile="local0"
quiescent=yes
negotiatetimeout=2

#Main http balancing
virtual=1.2.3.4:80
         service=http
         scheduler=wlc
         request="serverstatus.html"
         receive="NEW LB"
         checktype=5
         real=192.168.0.101:80 masq 3
         real=192.168.0.117:80 masq 4
         negotiatetimeout=1

.....

I've replaced real IP with 1.2.3.4 and cut plenty of real servers here 
and all rest of virtual sections. I'm clueless, config is mostly copied 
from old solution, running Debian 3.0, and probably older version of 
ldirectord/heartbeat. That Debian balancers are connecting to all real 
servers every 2 - 3 seconds, so it is rather not firewall or 
connectivity issue. I've tested firewall bu switching it off for few 
moments, didn't changed anything.

I would be glad for really any advice regarding this.

-- 
Regards
Piotr Gbyliczek
2nd Line Technican
ForLinux Ltd.


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.



<Prev in Thread] Current Thread [Next in Thread>
  • [lvs-users] Checkinterval issue., Piotr Gbyliczek <=