sapms.mon not starting

[Taurian]

 
 
Hi All,

We have configured SGeSAP and want to monitor SAP message server via the sapms.mon. We have made the following entries in file:

In package configuration file ascsC11.config:

SERVICE_NAME ascsC11ms

SERVICE_FAIL_FAST_ENABLED YES

SERVICE_HALT_TIMEOUT 5

In the control script ascsC11.control.script:

SERVICE_NAME[0]="ascsC11ms"

SERVICE_CMD[0]="/etc/cmcluster/C11/sapms.mon"

SERVICE_RESTART[0]=""

These are the two entries mentioned in the SGeSAP manual. But the issue we are facing is that the monitor is not starting. Everything else is working fine; cluster, package, SAN filesystems, SAP, all are running fine. Just the monitor script is not starting. What could be the issue? Are we forgetting any entry??

1. Why is the monitor not starting?

2. What name would be process name for the monitor?

Please help!

Regards,

Taurian.

[Taurian]

Hi All,

No replies?

Please help guyz. Is there a step we are missing?

We are using

- HP-UX 11.31

- Serviceguard version A.11.20.00.01

- SGeSAP B.05.10

Let me know if you need anymore details.

Regard,

Taurian.

[Taurian]

And we are using Legacy approach for package configuration.

Please help!

Regards,

Taurian.

[Matti_Kurkela]

1.)

After making any changes to the package configuration file, you must use the cmapplyconf command:

cmapplyconf -v -P ascsC11.config

 If this is not done, the configuration change will not be effective. If the package configuration file does not list any services for the package, the part of the control script that starts the services will never be called.

The real Serviceguard configuration file is a binary file: the text .config files can be thought of as "application forms to request a configuration change". You can edit the .config file as much as you want, but it won't change anything until you run cmapplyconf.

When you run cmapplyconf, Serviceguard will automatically make the configuration change on all cluster nodes in a synchronized fashion.

2.)

The process name should be either "/etc/cmcluster/C11/sapms.mon", or just "sapms.mon".

The SERVICE_NAME will be used in a "cmviewcl -v" listing, monitoring tools like Serviceguard Monitor and in Serviceguard SNMP status information.

[Taurian]

Hi Matti,

Thanks for your reply.

We have applied the configuration after every change. There are no messages relating to monitors in package log file.

The cmviewcl -v output shows:

Script_Parameters:
      ITEM               STATUS   MAX_RESTARTS  RESTARTS NAME
      Service            unknown  Unlimited     0        ascsC11ms

The status of the monitor is unknown. And when we kill the message service, the package continues to run, whereas it configured to fail. This means the monitor is not running or it is not working.

Regards,

Taurian.

[Matti_Kurkela]

Is the script /etc/cmcluster/C11/sapms.mon marked as executable?

("chmod a+rx /etc/cmcluster/C11/sapms.mon")

What happens if you run the script manually, as root, like this:

# cd /
# /etc/cmcluster/C11/sapms.mon

If the script is suitable for use as a Serviceguard service, it should start successfully and keep running until you interrupt it (e.g. by pressing Control-C). If it displays any error messages, find the causes of the errors and fix them.

Have you checked the package log for error messages related to this service?

[Taurian]

Hi Matti,

Thanks for your reply.

Yes, the sapms.mon has execute permissions:

-r-xr-xr-x   1 root       sys          27227 Feb 20 20:44 sapms.mon

There are no messages displayed in the package log file. That is one of our issues. If it would shows any errors, then we can troubleshoot them. But there is no acknowledgement of sapms.mon in the package log file.

Regards,

Taurian.

[Taurian]

Anyone??

Regards,

Taurian.

[Matti_Kurkela]

Did you try to run the sapms.mon manually, as I described in my previous post?

Please try it and see what happens. Does it produce any error messages? Does it die immediately? Does it keep running until you interrupt it?

[Taurian]

Hi Matti,

Thanks for your support.

I tried entering ". /etc/cmcluster/C11/sapms.mon" in the customer_run_cmds. This ran the monitor and messages started showing up in pkg log file. But I am not sure if this is the correct way to run the monitor.

Now I am faced with a new problem. The package status stays "starting" for ever.

Please guide me to the correct way to run the monitor. Thanks.

Regards,

Taurian.

[Taurian]

Hi All,

No body knows what's wrong with my scenario?

One very important question! Are the monitors that came with SGeSAP B.05.10 even supported for lecagy packages? Because I tried them for modular packages and they were running fine. Why are they not running for legacy packages?

Regards,

Taurian.