VMS system monitoring software for HP Openview.

[Gino Castoldi_2]

Hi,

Server: Alpha 4100, VMS 6.2

We are looking for a software product to monitor VMS systems. What we are looking for are ideas on what people are using to perform this function. Whatever the product is it should be able to integrate with HP's Openview/Operations product (ITO/OVO).

We are basically looking for a monitoring solution that can work with HP OVO.

10 points to any good answer or idea.
TIA, Gino

[labadie_1 · Accepted solution]

I am afraid you need Vms 7.1 to have such an agent, see

http://h18004.www1.hp.com/products/quickspecs/10450_div/10450_div.html

[Ian Miller.]

BMC Patrol will run on VMS V6.2 and feed events upwards. It's a bit nasty unix like though.

[Mike Naime]

Gino:

The short answer is "It can't be done"

We have/had the complete HPOV suite installed in our data centers. For monitoring the 200+ OpenVMS alphaservers we use BMC Patrol and Consoleworks.

We also use Consoleworks to monitor the 72 CLI ports on our HSG80 storage controllers.

Why repeat myself. :-)
For an earlier thread on this exact topic, go back to page 15 and look at this thread labeled MANAGING OPENVMS.
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=353508

Mike Naime.

[Wim Van den Wyngaert]

Gino,

I monitored at max about 200 servers and stations using simple DCL (4000 lines).

1 script that barely needs config but that monitored nodes 5.5, 6.2 and 7.x, using decnet 4 and 5, ucx 2.x,3.x,4.x,5.x.

I included a description of what it monitors. Some items are specific to this bank (but the script knows it).

Wim

[Wim Van den Wyngaert]

It didn't follow the internal http link

[Wim Van den Wyngaert]

Mike : what do you monitor via consoleworks ?

We use console manager but the scans are not active. What I'm missing is something that doesn't scan for known things but something that tells me when an unknown thing happens.

E.g. if the message "I don't like this my job. I am going to crash" appears on the console, that it is considered as unknown and thus an alarm must be given.

Wim

[labadie_1]

Wim, in the Report memory shortages of all kinds , I would add the npagedyn pool expansion failure. You can get it in sda, or even simpler:

$ mc agen$feedback
$ sea sys$system:agen$feedback.dat npagedyn_allocafail

The first line is the same as
$ @sys$update:autogen savparams savparams, but is shorter

[Wim Van den Wyngaert]

Labadie,

Thanks for the tip. I give an alarm when 50% of the max is used.

E.g. NPAGDYN 4, NPAGVIR 16 alarm at 8.

This alarm indicates that something is wrong or that the 4 is too low.

And autogen is too heavy for regular usage.

Wim (can't give you a point, sorry)

[labadie_1]

What you do should work nearly always, but I remember some stupid soft, trying to do directly a pool expansion from npagedyn beyond npagevir.

[labadie_1]

Wim, I read

Some tests may take lots of time. E.g. the test to check the highest versions. Writing a C program to do the time consuming things would be more performant.

Use DFU to find the files with version number above 30 000, it will be much faster

[Wim Van den Wyngaert]

Labadie,

One of the goals was not to use extra things. If I use DFU I need to install it everywhere. Also, since it is freeware it is normally not allowed by the bank.

God knows why they didn't include it in VMS anyway ? Some kind of "find" command as in U*** would be welcome.

Wim

[labadie_1]

DFU is good enough to come with Vms, at 7.3-2 or at another version, I do not remember, but I am sure of that. So it is not exactly what I call a freeware, it is written by Vms engineering, and supported.

[labadie_1]

You can install DFU, so that you have the help and so, but just copying dfu.exe to a directory is enough.

[Ian Miller.]

DFU is offically still unsupported freeware even though it is written by sombody at hp. I agree it should be included with VMS and I install it on every system but at present it is not.

Getting back to the original question you want software that monitors your systems and feeds events to HP OVO for AlphaVMS V6.2. As you have discovered your choice is very retricted due to the version of VMS you are running (9 years old). BMC Patrol is one solution I've seen work. You also need to monitor the consoles (can do this from a newer version of VMS). You can roll your own solution also but doing this comprehensively requires significant work. Implementation of any monitoring will be a non-trivial amount of work.

[Gino Castoldi_2]

Hi,

I am looking at a product from Heroix called
RoboMon for OpenVMS. Does anyone have any information,and/or opinions on this product?

10 points to any good answers or ideas.
TIA, Gino

[Ian Miller.]

ROBOMON - not used it personally but colleagues who have used have a good opinion of it. It is a native VMS app (which is a good thing) and will run on your version of VMS.

[Antoniov.]

Here HP page about Openview and some other software
http://h71000.www7.hp.com/openvms/system_management.html

Antonio Vigliotti

[Mobeen_1]

Gino,
I have used BMC Patrol for monitoring OpenVMS systems without any issues for few years now.

Its really easy to use and can be integrated without any issues.

You could download this from BMC website below
http://www.bmc.com

All you would need is ... install the BMC Patrol KM (knowledge Modules) for OpenVMS on your VMS servers and then you can install BMC Console on a XP desktop (running XP server).

The following url points to the products i am talking about
http://www.bmc.com/products/proddocview/0,2832,19052_19429_23134_7060,00.html

regards
Mobeen

[Mobeen_1]

Gino,
You can get all information about RoboMon for OpenVMS at the URL below

http://www.heroix.com/products/detail_robomon.htm

regards
Mobeen

[Wim Van den Wyngaert]

Mobeen,

Just hope that XP is there after a disaster.

Does BMC check quotas, tcp service limits, print queues, looping things (and the allowed ones), shadow set changes, file versions (32K), DTSS, global pages, device errors (and not the ones that are allowed such as recovered write errors), decnet transport limits, ... ?

Gino,

The doc of robomon is not very concreet. You'll have to ask a demo version.

[Wim Van den Wyngaert]

Mobeen,

Just checked your links. It seems BMC is doing most of the things. But how do you specify all the exceptions ?

E.g. certain images/users/process names do things that are forbidden to all others. And do you have to tell that for every node ?

Can you also add checks yourself ?

Wim

[labadie_1]

Wim

I am afraid you do things much more subtle that the standard tools

:-)

[Dale A. Marcy]

Robomon has been used here for several years and we are happy with it. You can customize it to run command procedures when events occur to tailor the notification you receive about the event.

[Gino Castoldi_2]

Hi,

Dale:

Are you running RoboMon with OVO or just NNM
by itself?

10 points to any good answers or ideas.
TIA, Gino