SNMP for Open VMS 7.2

[yoly]

Hi! i'm working on Open VMS 7.2 networking and i need to install an SNMP agent. I need to create a fonction to monitor two switches to know if they are alive or not, if not i will do a basculement... someone can help me please??? thanks a lot!!!!!

[Ian Miller.]

which TCP/IP product do you have installed and what version?

[Arch_Muthiah]

Yoly,

I would suggest that if you have tool CockpitMgr for OpenVMS, your job will become easy monitor netwok's all kind of activities. Otherwise we have to write TCPIP SNMP program and query the MIB database to get the network acitivity info.

Using this tool you can configure "SNMP agents" on network devices and SAN switches can be configured to transmit messages to well-known addresses in response to specific events. Then using CockpitMgr, we can monito SNMPgets info from SNMPTraps Listener.

To configure SNMP, we can do it easily using TCPIP$config and
1. TCPIP$SNMP_STARTUP.COM - Installs master and subagent images and runs TCPIP$SNMP_RUN.COM.
2. TCPIP$SNMP_RUN.COM - Starts the master agent and subagents.
3.TCPIP$EXTENSION_MIB_STARTUP.COM - Starts custom subagents.
4. TCPIP$VMS_SNMP_CONF.DAT- user editable config data file.
5. TCPIP$SNMP_CONF.DAT - Configuration data file used in the startup of the master agent and standard subagents.

And using SET CONFIG cmd, we can allwo master and subagent to communicate.
TCPIP> SET CONFIGURATION
SNMP /COMMUNITY="public" /TYPE=READ

TCPIP> SET CONFIGURATION SNMP /COMMUNITY="public" /FLAGS=SETS

Archunan

[yoly]

Hi!

Thanks for everybody who has replied me! but i'm new in this kind of system and i am a little lost :-(

I will explain another time, and better, my problem.

I have a system with two Alpha server with OpenVMS connected with two switchs (is a redounded system). I need that the supervisor (one of the alpha server's) goes to one of the switches and can take information of his state (alive or not). All that with SNMP.

I think this is :

- every switch must be a client,must have a MIB table with this information , isn't it?

- and the supervisor need to go to look this information every 1 second

- and finally the switches need to send a trap if a problem occurs (because the watchdog of the supervisor is only 1 second and it's a long time)

Maybe i'm not well explaining the problem, i know, but my english is not so good... ;-)

You can explain me a little more your ideas?

Thanks a lot!!!

PD. I put an attachment to know the topology of the network...

[Wim Van den Wyngaert]

Cockpit will cost you money.

If you don't have that, you can write a procedure in DCL. An extract of my procedure used for hub900 and sanswitches.

$ call do_snmp sanswcs2

do_snmp: subroutine
$ set noon
$ f="ver_proc_snmp_''p1'"
$ def/us sys$output 'f'.new
$ mc TCPIP$SNMP_REQUEST.EXE 'p1' public getnext -l -w 9 1.3.6.1.4.1
$ sea 'f'.new "No reply" , "Unexpected reply"
$ if $status .eq. %x00000001
$ then
$ call wms W "Check ''p1'. It is not responding properly to snmp."
$ del 'f'.new;*
$ exit 2 !generate abort
$ endif
$! 2.5.4.1.1.1 = SMNP authorization failure
$ sea 'f'.new /ma=no/out='f'.txt 1.3.6.1.4.1.36.2.18.11.2.5.4.1.1 , -
1.3.6.1.4.1.1588.2.1.1.1.1.1.0 , -
1.3.6.1.4.1.1588.2.1.1.1.6.2.1 , -
1.3.6.1.4.1.1588.2.1.1.1.1.22.1.4
$ del 'f'.new.*
$ if f$sea("''f'.txt") .eqs. "" then exit
$ diff 'f'.txt /par/mat=1
$ if $status .eq. %X006C8009
$ then
$ delete 'f'.txt;
$ exit
$ endif
$ call wms "E" "Possibly a failover problem : check device ''p1' in SNMP report (''f'.txt)"
$ exit
$ ENDSUBROUTINE

Wim

[Ian Miller.]

Some DCL build around the following program may be a way forward
http://h18000.www1.hp.com/support/asktima/communications/009B5362-5D9438CF-1C0186.html

[yoly]

Thanks all of you! i will begin to study your proposals...i will explain here my steps and the new questions that i will have in the following days!

Yoly