[lvs-users] sudden increase of ActiveConn
malte at snapscouts.de
Thu Nov 4 20:31:32 GMT 2010
try comparing the /etc/sysctl.conf "optimizations" .
For sure they did some adjustments there on the shorewall boxes (lowering tcp connection timeouts and so on)
Am 20.10.2010 um 19:40 schrieb L.S. Keijser:
> First of all let me say that this is a very vague and probably
> LVS-unrelated problem. I'm posting it anyway in the hope other people
> have run into this problem before and know the solution.
> We recently switched from an LVS setup using Shorewall for firewall
> rules management, to a setup using plain iptables. It's an LVS-NAT setup
> and the realservers that are behind the director are webservers. The
> only thing changed on the rs's is the default gateway which now points
> to the new director.
> What we notice is that there's suddenly an increase of ActiveConn.
> Before the switch it was around 400 at peak hours. Now it's close to
> 1400. If we point the rs's back to the old setup, the problem is gone.
> The old director is running CentOS 5.3 with kernel
> The new director is running CentOS 5.5 with kernel 2.6.18-194.8.1.el5xen
> Any tips would be most welcome.
> kind regards,
> Please read the documentation before posting - it's available at:
> LinuxVirtualServer.org mailing list - lvs-users at LinuxVirtualServer.org
> Send requests to lvs-users-request at LinuxVirtualServer.org
> or go to http://lists.graemef.net/mailman/listinfo/lvs-users
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 4227 bytes
Desc: not available
Url : http://lists.graemef.net/pipermail/lvs-users/attachments/20101105/f65b5894/attachment-0001.bin
More information about the lvs-users