Very Computer

Hi y'all;
I am trying to find any tools [commercial, or public-domain] that will assist
me in managing a collection of Unix servers and workstations.
We have a mix of Sun, Silicon Graphics, IBM RS6000, with a smattering of
HP/UX and Ultrix.

I am particularly interested in the following:-

1)      Being able to identify 'antisocial' processes that are hogging the
        CPU, using excessive amounts of memory or I/O.  [Some of our users
        write some inefficient programs!].

2]      Ability to set some thresholds for parameters such as memory
        utilisation, swapping, paging, I/O rate, number of processes,
        fullness of file-system; when these are exceeded some form of
        alert is generated. [I want to know when /tmp or /var is filling
        up *before* the users start complaining].

3]      Recording this info over a long period so it can be used in
        capacity-planning and system management.

It would be nice if there was some way to display this information, in
semi-real-time, using X-window based graphics. I have seen products for
the IBM mainframe/mini world, in which you can have each machine displayed
as an icon, normally green, but which goes amber or red to indicate that it
needs attention.

If you know of any products that may be of interest, please let me know!

Yours in anticipation...
--------------------------------------------------------------------------------
Peter J.M. Lucas          NERC Computer Services    Swindon    England



--------------------------------------------------------------------------------

: Hi y'all;
: I am trying to find any tools [commercial, or public-domain] that will assist
: me in managing a collection of Unix servers and workstations.
: We have a mix of Sun, Silicon Graphics, IBM RS6000, with a smattering of
: HP/UX and Ultrix.

: I am particularly interested in the following:-

: 1)    Being able to identify 'antisocial' processes that are hogging the
:       CPU, using excessive amounts of memory or I/O.  [Some of our users
:       write some inefficient programs!].

: 2]    Ability to set some thresholds for parameters such as memory
:       utilisation, swapping, paging, I/O rate, number of processes,
:       fullness of file-system; when these are exceeded some form of
:       alert is generated. [I want to know when /tmp or /var is filling
:       up *before* the users start complaining].

: 3]    Recording this info over a long period so it can be used in
:       capacity-planning and system management.

: It would be nice if there was some way to display this information, in
: semi-real-time, using X-window based graphics. I have seen products for
: the IBM mainframe/mini world, in which you can have each machine displayed
: as an icon, normally green, but which goes amber or red to indicate that it
: needs attention.

: If you know of any products that may be of interest, please let me know!

: Yours in anticipation...
: --------------------------------------------------------------------------------
: Peter J.M. Lucas          NERC Computer Services    Swindon    England



: --------------------------------------------------------------------------------
You might want to check out CA-Unicenter from Computer Associates.  It is a
cross-platform management tool that supposedly handles most of your list
above plus a lot.  I am just starting an evaluation of the product and so far
it looks good.

Ron Morley
Zenith Data Systems
--
Disclaimer: My opinions are my own...no one else wants them.
"The only two things a pirate'll     |"The least flexible component of any
 run for is money and public office" |system is the user" Lowell Jay Arthur
            - Yosemite Sam           |
"Never underestimate the power of human stupidity"  Lazarus Long

Quote:

>: 2]        Ability to set some thresholds for parameters such as memory
>:   utilisation, swapping, paging, I/O rate, number of processes,
>:   fullness of file-system; when these are exceeded some form of
>:   alert is generated. [I want to know when /tmp or /var is filling
>:   up *before* the users start complaining].

>: 3]        Recording this info over a long period so it can be used in
>:   capacity-planning and system management.

>: It would be nice if there was some way to display this information, in
>: semi-real-time, using X-window based graphics. I have seen products for
>: the IBM mainframe/mini world, in which you can have each machine displayed
>: as an icon, normally green, but which goes amber or red to indicate that it
>: needs attention.

[sigs and stuff about CA-Unicenter deleted]

---
                __                               ___  ___  __   ____
           /  /   \                             /    /  /  /   /
          /  /    /                            /__  /__/  /   /
   /     /  /    /                               / /  /  /   /
  /_____/  /____/                            ___/ /  / _/_  /____

John de la Garrigue                              ||  Phone:  619/686-5600

Science Applications International Corp. (SAIC)  ||  

My views are not necessarily those of my employer, the US government,
or anyone else for that matter.

What about that product that Tivoli has whose has a product that addresses
this?  Anybody out there used it yet?

Want a good semi-realtime performance monitor?  Check out HP's GlancePlus.
It's good, man.....

---
-----------------------------------------------------------------

Independent Contractor
(currently at SunSoft, Inc.)

You might want to look at EcoSystems' software.  They have versions for a number
of platforms (SunOS, SCO ...) for monitoring both operating system-level performance
AND database performance (Oracle, currently).  They can be reached at

        EcoSystems Software, Inc.
        10055 Miller Avenue, Suite 201
        Cupertino, CA  95014 USA

        Phone: (408) 252-3801 x109  Fax: (408) 252-3806

(Just a happy user)

        - John  
---
+---------------------------------------------------------------------+    

|  Supervisor, Network Operations       Phone: (302) 451-5200         |
|  Delmarva Power & Light Company       Fax:   (302) 451-5321         |
|  500 N. Wakefield Drive               NOC:   (800) 388-7076         |
|  Newark, DE 19714-6066                                              |
|  The opinions expressed are not those of Delmarva Power, simply the |
|  product of an over-active imagination...                           |
|  Time is Nature's way of preventing everything from happening at    |
|  once; it's not working too well here in NetOps!                    |
+---------------------------------------------------------------------+

Quote:>Hi y'all;
>I am trying to find any tools [commercial, or public-domain] that will assist
>me in managing a collection of Unix servers and workstations.
>We have a mix of Sun, Silicon Graphics, IBM RS6000, with a smattering of
>HP/UX and Ultrix.

It monitors: network traffic, disk usage, filesystem aging and other inter-
esting alerts.  Runs under X and makes really usage charts.

                                                - Gary Lin

                                                Resumix, Inc.

Brian

From the README:

rperf is similar to xmeter, or SunView perfmeter, but is intended for
use on ascii terminals or for logging.  rperf obtains system statistics
from one or more network hosts and prints them to the standard output
at periodic intervals, in a format similar to rup, vmstat, and
netstat.  rperf reports on cpu load, disk activity, network interface
activity, and virtual memory activity.

If installed as rup, the program prints the hostname, uptime, and load
average once, then exits.  A hard or symbolic link to rperf may be used
for the installation.