From: Pedro Cunha (pedro.cunha@decsis.pt)
Date: Mon Apr 29 2002 - 07:27:53 EDT
Hello all,
First of all thank's to everybody who spent some of their time answering to my question, too many to mention.
The most recomended tools where:
- Collect
- Sys-check
- Mrtg
- RRDTool
- Big Brother
I'm at this moment trying to use Big Brother to monitor and snmp/scripts + mrtg to collect statistics and create web pages.
All the replies are below.
cheers,
Pedro.
********************************************************************************************************************
Hi,
Use monitor, collect (good for all performance related
issues), top, also you can use sys_check for overall
system check.
cheers
Jay
********************************************************************************************************************
TOP is quick and easy for CPU
Monitor is OK.
Collect is comprehensive, best!
Kind regards
Lawrie Smith
Tru64 UNIX Systems Manager
Capita Technical Services
WestM alling
email: lawrie.smith@capita.co.uk< BR>Tel: 01732 877266
Mob: 0774 082 5468
********************************************************************************************************************
if you want to do some graphing, I believe Netscape's website has
some javascripting/HTML routines for doing bar charts.
George
********************************************************************************************************************
I wrote 2 C programs that runs every 2 seconds that log CPU/DISK usage to a rrdtool database.
Let me know if you need sample code. Very simple but hardcoded for number of disks.
Results are like:
********************************************************************************************************************
http://people.ee.ethz.ch/~oetiker/webtools/mrtg/
This is what we are using with good results.
HTH,
Chris Bryant
Unix Administrator
Dollar Rent A Car
(918) 669-3213
********************************************************************************************************************
sounds like what you want is mrtg + rrdtool (www.mrtg.org). It is
a very capable monitoring package for generating graphs and such.
-charlie
********************************************************************************************************************
Hi Pedro,
Let me start by saying that I work for Fortel, the company that I am going to recommend. You may consider me biased, but please don't discount SightLine on that basis alone. I have been involved in Performance Management and Capacity Planning for 15 years, and when I ran datacentres, SightLine was the product I chose for monitoring my systems.
SightLine has an Agent that runs on a server, and passes its data back to a PC for analysis. SightLine delivers between 1000 and 5000 different metrics every collection interval (Usually every 30 seconds) which makes it a real-time tool. Data can be displayed graphically to make it easy to understand. SightLine also has unbeatable performance analysis features, enabling you to set it up to give early warnings of problems before they occur. It also contains an expert knowledgebase, which can analyse your system and recommend changes you can make to improve performance.
SightLine can collect data from Tru64, OpenVMS, Solaris, Linux, NT, Win2K, HP-UX, AIX, Sequent Dynix, SCO, UnixWare, Stratus VOX, Unisys IX, Unisys IX, Oracle, EMC disk cabinets, InterSystems Cache', IIS, Exchange, SQL Server, Terminal Server, Tuxedo, OS/390, Cisco Routers and Switches, Apache, Firewalls, Network services (FTP, DNS, HTTP etc.) and a whole lot more. It will bring ALL of this data down to one single PC, where you can view, report and analyse all this data from a single place, and through one single piece of software.
SightLine also contains the only data correlation engine available in computer performance management tools today. This means that when a problem occurs, you can analyse everything that we are collecting data from, to find the root-cause of your performance problem. This means that if a router is the real cause, SightLine will show it up, whereas if you just examine the Tru64 box, everything may show up fine.
I hope this gives a taster,
Please visit our web site for more info.
If you need a demo or a chat, please call/e:mail me, and I'll ensure that a sales person gets back to you,
Best regards
Ashley
Ashley Shepherd FIAP MBCI
Consultant
Fortel (UK) Ltd
www.fortel.com
e:mail ashley.shepherd@fortel.com
Phone +44 (0)7768 934849
********************************************************************************************************************
You didn't specify whether you wanted statistics for alarms (event system) or performance statistics, so I'll mix it a bit.
There are various solutions, first those involving money :
BMC Patrol. http://www.bmc.com/ . Widely used and runs on just about all platforms. You can buy a web interface add-on, but it kinda sucks.
If you make a friendly call to BMC it's likely you can persuade them to send you a couple of demo CDs that last for 30 days.
BMC patrol can be really great if you take the time to learn it, and learn PSL scripting, but can also be a complete disaster if you don't take the time to learn it.
Performance Manager. Comes with the OS (and hence, you already have it). Start the daemon (pmgrd) and then the GUI (X-Windows). No web interface.
Then freeware/opensource :
RRDTool. http://www.rrdtool.org/ . It's really just a round robin database, but if you add the proper scripts that out data in a format that conform to the RRDTool standard, then it's fairly easy to get the statistics you want into it.
RRDTool comes with tools to extract the data as PNG graphics.
RRDTool is mostly a performance monitoring tool as it doesn't react in realtime on events.
Cricket. a SNMP based tool that allow you to store data in a RRDTool database. Can be used on Tru64 in conjunction with the Performance Manager SNMP daemon.
MRTG. http://www.mrtg.org/ . From the author of RRDTool, but older. Stores data in text files and mostly useful for SNMP
StatDB. http://statdb.dassic.com/ . Stores data in a database (MySQL or Oracle) and comes with a web interface, but can be used via ODBC (or plain SQL) too.
I won't comment whether it sucks or not, since the author is me.
There are likely more than these, but I don't remember more right now.
-- Un saludo / Venlig hilsen / Regards Jesper Frank Nemholt Unix System Manager Compaq Computer Corporation Phone : +34 699 419 171 E-Mail: JesperFrank.Nemholt@compaq.com ******************************************************************************************************************** We use MRTG for graphing statistics (I don't remember where to get MRTG from). We had to write some scripts to collect the statistics, and also a script to feed the 2 values to MRTG (it can graph any 2 numbers you throw at it). try MRTG / RRDTOOL http://people.ee.ethz.ch/~oetiker/webtools/mrtg/ <http://people.ee.ethz.ch/~oetiker/webtools/mrtg/> Bernhard Krieg ******************************************************************************************************************** ok, You can use collgui for graphs( which uses collect, but it needs perl n perlTk) Also you can you BigBrother which monitors your, cpu, mem, disks, again there is one more product sitescope,(not sure if it works on alpha but i have tested on solaris 2.7 ), cheers Jay. ******************************************************************************************************************** Hi Pedro, you should use mrtg. This software is free and originally written to monitor routers (and switches), but you also can use it to monitor any other kind of events, for example inside a tru64-system. Have a look at out web pages http://www.physnet.uni-hamburg.de/physnet/reports/mrtg/index.html and http://www.physnet.uni-hamburg.de/physnet/statistiken.htm - Axel _____________________________________________________ Axel G. Puskeppel PHYSnet-Rechenzentrum Jungiusstrasse 9 D-20355 Hamburg Tel.: +49 (0)40 42838 2504 Fax: +49 (0)40 42838 2503 eMail: puskeppel@physnet.uni-hamburg.de http://www.physnet.uni-hamburg.de/physnet ******************************************************************************************************************** go to www.freshmeat.net and search for network monitoring tools. MRTG is an obvious candidate. Good luck, -- Steve Herber herber@thing.com work: 206-261-0307 Systems Engineer, AMCIS, UoW home: 425-454-2399 ******************************************************************************************************************** Ignoring the tools you have to pay for, some of the common tools are: *stat; iostat, vmstat, netstat, etc. These are included in the base system. They just display text data, but suitable filtering software could probably generate graphs. Monitor - This collects data in a documented format, which can be processed as you see fit. The data files can also be replayed using the same interface (Curses based) that it uses to display live data. Collect - I think it has some of its own graph generation capabilities. Whether it is suitable for direct inclusion into a web page is something I can't comment on. ******************************************************************************************************************** Have you tried Collect. The collect utility it's includes in version 4.0 G and up but to use the graphic you need the collgui which you can download for free in the compaq site. Regards, Iwan ******************************************************************************************************************** Pedro, i am looking into a product call MRTG it produces very nice web based graphical output. It uses snmp to gather information. http://www.mrtg.org Regards ******************************************************************************************************************** Dear Pedro, I use collect to gather information for my machines and collgui to display results. collect and collgui are from Compaq and can be found at : http://www.tru64unix.compaq.com/collect/ Regards, Cetin Haftaoglu ******************************************************************************************************************** Pedro Cunha (Systems Engineer) Decsis - Sistemas de Informação, Lda. Rua Pedro Hispano, 1329 4250-368 Porto Portugal Tel: +351 228349310 Fax: +351 228349319
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:48:39 EDT