LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: timeout question

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: timeout question
From: Graeme Fowler <graeme@xxxxxxxxxxx>
Date: Thu, 13 Apr 2006 09:38:31 +0100
Hi Leon

On 13/04/2006 09:26, Leon Keijser wrote:
Not yet, no. Since it didn't matter (according to the sysadmin) on which
server the clients would connect to after a disconnect. I can turn it on.
Does it matter which persistence value i add for it?

Thinking about it, this is probably unnecessary and could cause you problems if a server "goes away", so we'll file that idea - was me thinking out loud ;-)

Exactly. Almost the entire time the client is idle with no traffic going
either way. The system is for looking up patient data etc, and is usually
idle when there's nobody at the counter.

Right then - I have no idea what sort of client or server you're using, so this may (or may not!) be an option... the Linux telnet daemon (amongst many others) normally starts up with TCP keepalives enabled. This can be overriddden (quoting the man page):

-n   Disable TCP keep-alives.  Normally telnetd enables the TCP keep-
     alive  mechanism  to  probe  connections that have been idle for
     some period of time to determine if the client is  still  there,
     so  that idle connections from machines that have crashed or can
     no longer be reached may be cleaned up.

PuTTY (the Telnet/SSH client) can send keepalive packets in the other direction (probably using the telnet NOP command, I think) to achieve the same end.

Either way, this means that there are 'null' packets - but they are packets - being transferred between client and server, and that _should_ be enough to keep your sessions open.

Of course, this entirely depends upon the client and server being used...

Graeme

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