[CALUG] Linux resource usage

[Rajiv Gunja]

Depending on what type of Linux system you are monitoring, the type/need
of/for tools varies.

If you are going to monitor a destop system, then there are a multitude of
karamba/superkaramba/gdesklets available which gives you the network,
hard-disk, cpu information every few seconds, some easily changed as they
are scripts, others need some tinkering using python.

If you are monitoring a server, then you will need to script the things you
would like to monitor using tools either already present or installing them
additionally. The easy way is to either program you script into a desktop
applet or send it to an html file which is updated at a said interval.

Tools you would need are:

   1. sensors ( to monitor the voltage and temperature of the cpu, fan and
   motherboard)
   2. hddtemp (to monitor the temperature of your hard-disks)
   3. hwinfo (to keep your hardware configuration available when server
   hardware fails, good for disk serial number and firmware version)
   4. dmidecode (to keep track of your hardware serial number, firmware
   version and BIOS info - useful when calling vendors)
   5. iostat, mpstat, sar, top and vmstat to get the statistics or the
   current state of the live system.
   6. smartmontools (useful to monitor disk degradation / failures)


Hope this helps.

-GGR

--
Rajiv G Gunja
Blog: http://ossrocks.blogspot.com


2009/1/26 Jason Dixon <jason at dixongroup.net>

> On Mon, Jan 26, 2009 at 04:21:19PM -0500, Jason C. Miller wrote:
> >
> > On Mon, 26 Jan 2009, Jason Dixon wrote:
> >
> >> On Mon, Jan 26, 2009 at 04:14:41PM -0500, Jason C. Miller wrote:
> >>> Hey all.  :)
> >>>
> >>> I'm looking for a good way to monitor resource usage for a linux
> machine.
> >>> Redirecting 'top' to a file only goes so far and I'm NOT looking to do
> >>> something as drastic as SNMP.  Here's what I'm looking for
> >>> in a solution:
> >>>
> >>> 1. No extra software purchase/download necessary
> >>> 2. If software download is mandatory, that it be extremely minimalistic
> >>> 3. Uses the default linux toolbox of utilities
> >>> 4. Something thorough (seperate CPUs, memory, I/O)
> >>>
> >>> It could be that something simple already exists and that I've missed
> it.
> >>> If you know of a cool utility, script, or method that I can use to
> >>> accomplish this, please chime in.
> >>
> >> There's nothing drastic about SNMP.  It's very simple once you have the
> >> basic command string and was designed for what you want.
> >>
> > True.  But I would like some fine-grained statistics (per second, if
> > needed) without muddying the network.  This just skews the I/O results
> > anyway.
>
> You can pull SNMP queries as often you like.  SNMP != rrdtool.  It only
> skews your results if you're querying from another system.  The same is
> true for *any* administrative traffic you'd use to monitor the system
> remotely.
>
> --
> Jason Dixon
> DixonGroup Consulting
> http://www.dixongroup.net/
>
> _______________________________________________
> CALUG mailing list
> CALUG at unknownlamer.org
> http://lists.unknownlamer.org/listinfo/calug
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.unknownlamer.org/pipermail/calug/attachments/20090126/3eda1e6c/attachment.htm