LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: [lvs-users] Kernel 2.6.35 and 100% S.I. CPU Time

To: John Sullivan <jslvs@xxxxxxxxxxxxxxxxxxxx>
Subject: Re: [lvs-users] Kernel 2.6.35 and 100% S.I. CPU Time
Cc: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
From: Julian Anastasov <ja@xxxxxx>
Date: Wed, 6 Oct 2010 11:06:12 +0300 (EEST)

        Hello,

On Mon, 27 Sep 2010, John Sullivan wrote:

I backed out my patch and tried this instead.

It does appear to fix the 100% SI problem, but creates a new
one. Under my test load (fire 100 simultaneous HTTPS requests using
wget at the load-balanced address), all the wgets hang, and netstat
on the backup server shows all 100 connections in this state:

Proto Recv-Q Send-Q Local Address Foreign Address State       PID/Program name
tcp      113      0 lb-addr:443   ext-host:49914  ESTABLISHED -


It looks to me like packets on those connections are now just being
discarded (even when they shouldn't be).

        Sorry for the delay. I'm attaching new patch for
testing. You should apply it on backup server. For your
setup only the last hunk in patch is needed. One of the
changes will keep the forwarding methods as specified,
may be you will not see "Local" real servers. If this is
a problem for you let me know. You can test then only
the last change in ip_vs_dr_xmit. Attached patch is for
2.6.35. It allows backup server after being used as
DR/TUN server in master to forward locally the packets
for connections created after sync.

Regards

--
Julian Anastasov <ja@xxxxxx>

Attachment: local_rs3.diff
Description: Support backup as DR in master

_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/

LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
or go to http://lists.graemef.net/mailman/listinfo/lvs-users
<Prev in Thread] Current Thread [Next in Thread>