On a fairly 'idle' Linux for S/390 machine (running only an FTP server
and the Hobbit client), ESAMON measures the CPU utilization at about .1%
- .2% of a z890 IFL. And that is only when the agent is active, the
rest of the time the machine reports much less (or no) utilization. The
agent wakes up every five minutes and reports information back to the
server (CPU, memory, disk utilization and more).
The server is, of course, somewhat more active depending upon when
messages come in and when it runs it's own tests.
Even though Hobbit reports and tracks CPU utilization and other
performance metrics, it is *not* intended as a performance monitor. It
does this to be able to report when thresholds are exceeded.
I can't speak to Nagios, I've never actually used it.
Barton Robinson wrote:
One of the most critical issues with Linux on "z" is the
overhead of the infrastructure. If you plan on running a hundred
linux servers, multiple the cost of the agents required by the
infrastructure times 100 and then ask if this is a workable
solution.
NETSNMP runs at between .1 and .3% of a processor on most
servers - it is used primarly for performance data collection
and accounting purposes, it is also used for operational alerts
on file systems, disk space, looping processes, swapping, and server
availability,
There are thousands of linux servers running netsnmp today
on the "z" platform, probably orders of magnitude more than any
other agent because of it's low overhead and high functionality.
Any idea what the cost is of operating Nagios, or Hobbit is, with
there relatively low functionality? I haven't heard of anyone
running these on systems with many servers, nor have i heard
of anyone actually running the 892 pound leviathan on this platform
in production anywhere.
Date: Fri, 3 Mar 2006 08:57:07 -0600
From: Rich Smrcina <[EMAIL PROTECTED]>
Additionally, I've integrated monitoring VM and VSE into Hobbit (disk
space, running processes [virtual machines or jobs], CPU, Paging, etc).
There is a client for z/OS and OS/390, but since I don't have access
to z/OS (nor am I well versed in z/OS) I don't have any examples.
I would guess the same could be done for Nagios with modifications.
David Boyes wrote:
Anyone using this to monitor Linux under VM on a large scale basis?
Seems like a pretty heavy agent (java based) memory wise.
It is a 892 pound leviathan (I won't recap the long discussion I had
with the Director development team about *why* this was bad). If you
have heavy investment in Director on Intel or PPC and you're using
Tivoli Provisioning Mgr or Intelligent Orchestrator it may make sense to
integrate with the automated provisioning stuff, but the footprint of
the agent is *enormous* for simple health-check monitoring.
"If you can't measure it, I'm Just NOT interested!"(tm)
/************************************************************/
Barton Robinson - CBW Internet: [EMAIL PROTECTED]
Velocity Software, Inc Mailing Address:
196-D Castro Street P.O. Box 390640
Mountain View, CA 94041 Mountain View, CA 94039-0640
VM Performance Hotline: 650-964-8867
Fax: 650-964-9012 Web Page: WWW.VELOCITY-SOFTWARE.COM
/************************************************************/
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
--
Rich Smrcina
VM Assist, Inc.
Main: (262)392-2026
Cell: (414)491-6001
Ans Service: (360)715-2467
rich.smrcina at vmassist.com
Catch the WAVV! http://www.wavv.org
WAVV 2006 - Chattanooga, TN - April 7-11, 2006
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
