1833845 Members
2229 Online
110063 Solutions
New Discussion

Service Guard

 
SOLVED
Go to solution
eddier
Frequent Advisor

Service Guard

I have some doubt...

I have an NNM server running in HP UX cluster.
I have nnm.mon, .cntl and conf files for SG cluster given by HP..

Whenever any of my services get failed the SG packages stops all NNM application services, unmounting /etc/opt/OV/share and var/opt/OV/share and removing the VIP.

But it is not get shifted to the failover server.

Anyone can help me in this...

Is it possible to get shifted to the Failover server when NNM services got failed

If someone need the scripts i ll give my packages scripts custimozed for NNM application...

regards
Eddi
10 REPLIES 10
melvyn burnard
Honored Contributor
Solution

Re: Service Guard

so what does cmviewcl -v show for this package? Is switching enabled?

Post the output here
My house is the bank's, my money the wife's, But my opinions belong to me, not HP!
TaaKie
Occasional Advisor

Re: Service Guard

perform a "cmmodpkg -e " to enable switching

Default switching is disabled when you manuallt switch the package.


Hope it helps,
TaaKie
eddier
Frequent Advisor

Re: Service Guard

Thanks for both of u...

I have attached two commands in a notepad. Plz have a look.

cmviewcl -v
cmviewcl

hi
i accept the chmodpkg command. We did the testing using this command only.

But my question is should i run this chmodpkg manually wont it be automatic ?

cheers
eddi
eddier
Frequent Advisor

Re: Service Guard



In that cmviewcl output u can see nnm01 package. That is the nnm package. plz ignore ovpi01 package output...

Then can u explain me the differece of these two..
********************************************
1.
Node_Switching_Parameters:
NODE_TYPE STATUS SWITCHING NAME
Primary up enabled cimph002(current)
Alternate up enabled cimph001

********************************************
2.
PACKAGE STATUS STATE AUTO_RUN NODE
nnm01 up running disabled cimph002

********************************************

By default this Auto_run is disabled state when we run this chmodpks this state changes to enabled. As soon as i execute the chmodpkg command it is shifting to failover server. why can't that Auto_run be enabled forever. so that whenever my NNM service get stopped it get shifted to Failover server automatically.

cheers

Eddi
melvyn burnard
Honored Contributor

Re: Service Guard


Then can u explain me the differece of these two..
********************************************
1.
Node_Switching_Parameters:
NODE_TYPE STATUS SWITCHING NAME
Primary up enabled cimph002(current)
Alternate up enabled cimph001

********************************************
2.
PACKAGE STATUS STATE AUTO_RUN NODE
nnm01 up running disabled cimph002


Output 1 shows what is called LOCAL switching, or whether the node listed is allowed to run the package locally

Output 2 shows what is called GLOBAL switching, or whether the package is allowed to switch between nodes in the cluster.
You need to check whether your package is configured to automatically switch, and if not chnage this, or if it is configured to switch, then you need to look at why it was disabled.
Check the package configuration file for the parameter AUTO_RUN and see if it is set to YES


My house is the bank's, my money the wife's, But my opinions belong to me, not HP!
Bharat Katkar
Honored Contributor

Re: Service Guard

Hi Eddi,

You need to edit /etc/cmcluster/ovpi01/ovpi01.conf file and set AUTO_RUN attribute to YES.
Save the file and then you need to do "cmcheckconf" and apply the configuration using "cmapplyconf".

Hope taht helps.
Regards,

You need to know a lot to actually know how little you know
eddier
Frequent Advisor

Re: Service Guard

The AUTO_RUN already set to YES only.

We have did that commands u have listed.

I am attaching my nnm01.conf file along with this.

FYI it is nnm01.conf file and not ovpi01.conf. Plz refer previous messages.

cheers

Eddi
melvyn burnard
Honored Contributor

Re: Service Guard

Ok, this basically then indicates that at some time prior to your package wanting to switch, an action was performed that disabled global switching. This could have been an error condition, a manual package halt or switch are examples.
You would need to set everything back to normal, and ensure all switching is enabled, then observe what happens during the failure and check the syslogs/package logs to see if it did switch or if it failed, why.
My house is the bank's, my money the wife's, But my opinions belong to me, not HP!
eddier
Frequent Advisor

Re: Service Guard

i checked the syslogs. It automatically takes "Disabling Global switching". Then packages get stopped.

Any more suggestions

regards
eddi
Bharat Katkar
Honored Contributor

Re: Service Guard

hi,
Can you attach or go through package log files which is kept in /etc/cmcluster//.

Hope you get some clue there.

Regards,
You need to know a lot to actually know how little you know