LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: max persistance time?

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: max persistance time?
From: Joseph Mack NA3T <jmack@xxxxxxxx>
Date: Thu, 31 Aug 2006 13:42:11 -0700 (PDT)
On Thu, 31 Aug 2006, Joseph T. Duncan wrote:

2 bad things can happen..

applications can be running full bore (think long batch type jobs.. and use a 100% of a cpu,

:-(

last write wins..

:-(

hopefully the keepalive settings will mitigate this ;)

mitigate, but it's not going to solve it. All clients with a dirty disconnect are going to leave a 100% CPU unattached job. The keepalives are only going to keep alive the sessions with a clean disconnect.

Can you differentiate the sessions you don't want from the ones that you want to stay idle?

not really.. on campus we have some wireless labs and wireless access that are behind a nat-proxy.

OK

also I dont know how to test for an idle connection on the director...

not much hope there unless you can do L7 (I don't know what L7 will look for, the absence of a clean disconnect?). However as Malcom says in the L7 section of the HOWTO, L7 should be done by the application, that's what the application is for.

there is a windows management tool that reports idle time but I am not aware of mib/snmp way to export that information

so you have two problems

o keeping up idle (or backgrounded) sessions that arise from a clean disconnect. You should be able to keep these open for any large time (eg weeks) if they aren't using any significant resources.

o killing 100% CPU sessions from a dirty disconnect.

does the application know whether the client has done a clean disconnect or not? I assume no or else you wouldn't be posting at all.

What does the app vendor say about this problem?

How do you handle the problem when there is no LVS?

I could make the active but idle timeout on the real servers much lower. but that would lead to unhappy proffessors that leave themselfs logged in overnight on weekdays.

I leave some of my xterms open for months. A professor should be able to do the same thing.

Joe

--
Joseph Mack NA3T EME(B,D), FM05lw North Carolina
jmack (at) wm7d (dot) net - azimuthal equidistant map
generator at http://www.wm7d.net/azproj.shtml
Homepage http://www.austintek.com/ It's GNU/Linux!

<Prev in Thread] Current Thread [Next in Thread>