HACMP clstop doesn't seem to work

[svnhunix]

Hi

I'm trying to write a perl script that stops the cluster on one node. So I'm looking foor the right commandline input to stop the cluster.

I tried:
/usr/es/sbin/cluster/sbin/cl_clstop -cspoc "-n Twin2HACMP1" -fyNs
and
/usr/es/sbin/cluster/utilities/clstop -f -y -s -N -S

As I understand the manpage both of them should stop the cluster and the clusterservices.

When I do cldump on the second node I get the output that both nodes are still online and that the application is unmanaged (wich is normal). On the first node (the one I want to stop the cluster) all services are still running:
Twin2HACMP1.axiroot:/home/users/axiroot> /usr/es/sbin/cluster/sbin/cl_clstop -cspoc "-n Twin2HACMP1" -fyNs
Twin2HACMP1: Sep 5 2008 11:25:11 /usr/es/sbin/cluster/utilities/clstop: called with flags -f -y -N -s
Twin2HACMP1.axiroot:/home/users/axiroot> lssrc -a |egrep "svcs|ES"
clcomdES clcomdES 344268 active
clstrmgrES cluster 266400 active
topsvcs topsvcs 245860 active
grpsvcs grpsvcs 413856 active
emsvcs emsvcs 364672 active
emaixos emsvcs 385072 active
clinfoES cluster 422102 active
grpglsm grpsvcs inoperative

These are the last lines in hacmp.out:
[...]
Sep 5 11:25:16 EVENT COMPLETED: node_down_complete Twin2HACMP1 forced 0

:check_for_site_down_complete[+54] [[ high = high ]]
:check_for_site_down_complete[+54] version=1.4
:check_for_site_down_complete[+55] :check_for_site_down_complete[+55] cl_get_path
HA_DIR=es
:check_for_site_down_complete[+57] STATUS=0
:check_for_site_down_complete[+59] set +u
:check_for_site_down_complete[+61] [ ]
:check_for_site_down_complete[+72] exit 0

Every command was runned as user root.

Can anyone help me with determining why the cluster services don't stop?



Thx

 

[DSMARWAY]

Hi

What version of HACMP are your running ?

When you stop the cluster does your application/ volume groups/ filesystems get stopped/umounted ?

 

[svnhunix]

Hi

it is v5.4

I've scratched the cluster config this morning and configured it again. Now the problem seems to be solved. I can start and stop the clustersoftware normally again.

Thx for the reply though