Approach 1:How to remove service group on the cluster with no interruption to the production application by restarting HAD
- Freeze all the service groups as a precaution.
# haconf -makerw
# hagrp -freeze -persistent
# haconf -dump -makero
- Stop HAD (with -force that will leave the actual resources running).
#hastop -all-force
3. Make a backup copy of the /etc/VRTSvcs/conf/config/main.cf file.
cp /etc/VRTSvcs/conf/config/main.cf /etc/VRTSvcs/conf/config/main.cf_bu
4. Edit and remove the service group from /etc/VRTSvcs/conf/config/main.cf
.
#vi /etc/VRTSvcs/conf/config/main.cf
Remove SG from main.cf.
5. Copy the configuration file to other nodes of the Cluster
# rcp /etc/VRTSvcs/conf/config/main.cf node2:/etc/VRTSvcs/conf/config/main.cf
6. Start VCS on all nodes
#hastart
Approach 2: Without restarting HAD
1. Permit dynamic changes to the VCS configuration file.
#haconf -makerw
2. freeze all theservice groups.
#hagrp -freeze -persistent
3. Remove any Service Group level dependencies.
#hagrp -unlink
4. Remove resources from the service groups one by one :
#hares -delete
5. Delete the service group :
#hagrp -delete
6. Verify that SG have been removed from VCS and unfreeze the other SGs which you want to :
#hagrp -unfreeze -persistent
7. Save the configuration
#haconf -dump -makero