topic Re: Monitoring console manager lines in Operating System - OpenVMS

Monitoring console manager lines

Wim Van den Wyngaert — Wed, 04 Jan 2006 05:46:29 GMT

Hi,

I have a number of server consoles connected to a decserver. I use console manager to reach these consoles. Also hub900 and disk controllers are connected.

Since VMS rarely requires console access I rarely access these consoles. But I want to be sure to find a working one when it is necessary. And sometimes it isn't working.
E.g. decserver gone, cable gone, port blocked, etc.

How can I monitor all these things (in dcl from a VMS node) ? There is no message in the console manager log files ! And not all consoles frequently display messages.

Wim

Re: Monitoring console manager lines

Ian Miller. — Wed, 04 Jan 2006 06:32:50 GMT

the way I have previously done this for monitored systems is to have something that does send a message to the console every 5 mins . The console monitoring software raises an alarm if it sees no activity.

Another posibility is to see if the DECserver will send an snmp trap when it sees a signal change on the port and to receive this trap. I think the VMS Cockpit Manager software (excellent software from hp Belguim) does this.

Re: Monitoring console manager lines

Wim Van den Wyngaert — Wed, 04 Jan 2006 07:03:10 GMT

Ian,

But what about controllers and hub900. How do you force a message on them ?

I don't consider SNMP traps because they could be missed (yet another bad industrie standard imo, it should trap until a confirmation is received).

Also, how do you raise the alarm ? In console manager itself ? How do you get it to another monitoring system ?

Wim
(junior in console manager)

Re: Monitoring console manager lines

Robert_Boyd — Wed, 04 Jan 2006 12:29:55 GMT

Wim,

What kind of disk controllers are you running?

I set up some DCL and scripts to ping disk controllers in the past. I might be able to help with that.

I don't currently have access to a console manager system, but I worked with it from 1994 to 2003, almost every day.

I had my entire setup making sure that the connections were intact on a pretty steady basis. At least every 30 or 60 minutes.

Robert

Re: Monitoring console manager lines

Wim Van den Wyngaert — Thu, 05 Jan 2006 01:47:07 GMT

Robert,

Even once a day is enough for me.
I have all kinds of controllers but those that will stay a while are HS1CP, HSG80 and HSZ70.

Wim

Re: Monitoring console manager lines

Lokesh_2 — Thu, 05 Jan 2006 01:58:53 GMT

Hi Wim,

We use a script to monitor console line everyday. The script check the console lines (CONS STAT/ALL), and also use console dialog feature to see if we find any response (like USERNAME) at the console to check that console is working fine.

HTH,
Thanks & regards,
Lokesh

Re: Monitoring console manager lines

Wim Van den Wyngaert — Thu, 05 Jan 2006 02:09:57 GMT

Hi Lokesh,

That's new for me. I will read the manuals and try it out. So it was worth asking it.

Wim

Re: Monitoring console manager lines

Wim Van den Wyngaert — Thu, 05 Jan 2006 03:38:12 GMT

Did some testing with it.

console dialog /file=wim.lis

and wim.lis containing

rate 2
send " "
wait 3 ""

1. Decserver lat port can not be reached.

-> Can't open console, status X1000006A

2. Decserver LAT port can be reached but cable between decserver and server is not functioning or not plugged in.

--> Dialog error : WAIT timed out, status X!0000074

3. Connection OK.

Status X10000065

Tested a working hub900, hsz70, hs1cp. VMS node at boot prompt, VMS node logged off, VMS node logged in. All work.

Did not work : node logged in doing a wait.

Good enough for me.

Thanks a lot Lokesh

Wim

Re: Monitoring console manager lines

Wim Van den Wyngaert — Thu, 05 Jan 2006 04:40:49 GMT

Here is a template of how to implement the verification.

Had to add UNLOCK to get it properly working on all nodes. This does not do UNLOCK as in console manager but clears the console for usage.

$ set noon
$ create wim2.lis
rate 5
unlock
send " "
wait 10 ""
$ def/us sys$output wim.lis
$ cons stat/all
$ open/read x wim.lis
$ read/end=e x x_rec
$r:
$ read/end=e x x_rec
$ nod=f$extr(3,16,x_rec)
$ console dia /fi=wim2.lis/quiet 'nod'
$ if $status .ne. %X10000065 then say "''nod' NOK"
$ goto r
$e:
$ close x

Re: Monitoring console manager lines

Wim Van den Wyngaert — Thu, 05 Jan 2006 10:03:09 GMT

Update.

1 node that was upgraded from 6.21h3 to 7.3 had a console manager working badly. The dialoge always timed out. Reinstall console manager and problem solved.

1 node (Alphastation 200) on bootprompt only reacted from time to time. Power cycle solved the problem.

Note that output is generated in console manager log (so if you do it every 5 minutes you may run out of space).

Wim

Re: Monitoring console manager lines

Robert_Boyd — Thu, 05 Jan 2006 14:39:41 GMT

Wim,

For HSx type controllers I wound up sending the text string "! ping" over the wire. And I looked for the string to echo back. If the wire is good and the controller is in the normal prompt mode (which is the way it should be left) then all is good.

I'll see if I can scare up some sample DCL for you and post it here.

Robert

Re: Monitoring console manager lines

Wim Van den Wyngaert — Fri, 06 Jan 2006 01:56:50 GMT

No problems with the controllers until now.
I did a check of the line every 10 minutes all night long.

But the station I had to reboot reported 2 times "Unable to connect" and a GS160 node 1 time "wait timeout".

I extracted console log with "/time" sand found that the GS160 node reacted to slow, so I increased the timeout from 5 to 10 seconds.

For the station, nothing was seen. So, I will do a send twice now.

Wim