[aklug] Re: Reasonable average server load?

From: Arthur Corliss <acorliss@nevaeh-linux.org>
Date: Thu Jun 24 2010 - 10:37:39 AKDT

On Thu, 24 Jun 2010, adam bultman wrote:

> Yes, but the VMWare folks - at least the ones I talked to via support -
> don't suggest you do that. They say that it should handle memory just
> fine. That being said, I *did* go back and make changes to the ESX
> servers and the VM so that it's RAM never gets swapped out again. And
> the munin graphs let me see right away if the ESX server is doing
> anything dumb.

If the vendor failed to ask about the operating profiles of your guests I'd
have to say they're not only wrong, he's incompetent. Allowing guest RAM
pages to be swapped out on the host would only be appropriate for specific
work loads. Say a redundant/backup server who's predominantly idle. For
any server with an active work load you're just asking for trouble by not
pinning memory.

> ESX has *some* limited functionality (vmstat, but no iostat or sar) that
> you can use to see how it is doing via the command line, but it seems
> VMWare really prefers you to use the windows client they have, and then
> just look at the graphs. There's other fancy-pants VMWare utilities you
> can use, but parsing the output is horrid.

I'm not surprised. Sounds like a good opportunity to transfer over some
statically-compiled binaries, though. ;-)

         --Arthur Corliss
           Live Free or Die
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Thu Jun 24 10:37:47 2010

This archive was generated by hypermail 2.1.8 : Thu Jun 24 2010 - 10:37:47 AKDT