HACMP does not start db2 after failover (db2nodes not getting modified by hacmp)

hi,

when I do a failover, hacmp always starts db2 but recently it fails to start db2..noticed the issue is db2nodes.cfg is not modified by hacmp and is still showing primary node..manually changed the node name to secondary after which db2 started immediately..unable to figure out why hacmp is not doing it automatically as it should.. I have checked permissions, all configuration I can think of and even compared to working ones but could not find any diff...any help on the issue is greatly appreciated.

thanks,
g rob

Did you check the hacmp-logs and errpt?

try to run your application server script manually and check where it fails

for db2nodes.cfg, I use the following lines in my Scripts:

         
                echo "0 $(hostname) 0" > /db2/db2${HASID}/sqllib/db2nodes.cfg 
                chown db2${HASID}:db${HASID}adm /db2/db2${HASID}/sqllib/db2nodes.cfg
                chmod 444 /db2/db2${HASID}/sqllib/db2nodes.cfg  

As I understand, modification of db2nodes.cfg file is the responsiblity of the DB2 start script. So, if that is not happening then something is wrong in the DB script and not an issue with HACMP configuration itself?

In simple, HACMP is not directly responsible for modifying this file with the correct nodename..right? instead it should be the DB2 start script(which is initiated by HACMP) that should be doing it?

the modification is not built in, so you manually have to make it

there must be an application server configured on your cluster, search for that script, and try to add my lines, with your paths of course

db2 checks, if the local hostname is equal with the hostname in db2nodes.cfg

if not, the database wont come up

maybe there is another method to achive this, but simply editing db2nodes.cfg worked fine for me for years now

changing the hostname in the cluster is another solution to solve this problem, but this may lead to problems with other applications