[lvs-users] Large clusters and slow realserver checking

Malcolm Turnbull malcolm at loadbalancer.org
Thu Sep 24 15:04:29 BST 2009


We were just about to start looking into the fork method for ldirectord (we
currently use an older version....)
Loadbalancer.org Ltd would be happy to sponsor a Perl hacker to increase the
performance of the ldirectord health checks,
as we don't have the right skills in house at the moment.

If anyone is interested by all means contact me off list, everything would
be contributed back to Horms/Ldirectord of course....

2009/9/24 Anthony Sturchio <asturchio at ambassadorbookstores.com>

> My company has a large number of secure web applications that we have
> running on an LVS cluster.  There are about 200 real IP's (for 200
> different
> domains / SSL certificates) and we have 5 different realservers in the mix.
> Since we are serving http and https (80 and 443) this works out to be 2,000
> realserver entries that ldirectord has to go through.  Obviously this takes
> some time.  We have seen it take up to 15 minutes to expire a downed node,
> or to reinstate a realserver once we bring it back up, of course depending
> on how far along the list ldirectord is.  Using the forking option is not
> possible since spawning a high number of processes simultaneously brings
> the
> load balancer to its knees.


Malcolm Turnbull.

Loadbalancer.org Ltd.
Phone: +44 (0)870 443 8779

More information about the lvs-users mailing list