Cleaning an incomplete Db2 pureScale instance drop (Linux)

When the db2idrop command has failed to completely drop the instance, you must complete the clean up of a Db2 pureScale instance.

To clean up an incomplete Db2 pureScale instance drop:

1. Stop and disable the Pacemaker and Corosync services:

   systemctl stop pacemaker
   systemctl stop corosync
   systemctl disable pacemaker
   systemctl disable corosync

   1. Delete /etc/corosync/corosync.conf
   2. Delete all files under /var/lib/pacemaker/cib/*
2. Clean up the Db2 cluster file system.
   For more information about cleaning up the Db2 cluster file system, see Manually cleaning a DB2® managed clustered file system.
3. Clean up $HOME/sqllib for the instance user on all hosts.
4. Clean up the entries in /etc/services for the Db2 pureScale instance user.
5. Remove instance signature

   installed_path/instance/db2iset -d instance_name

6. Remove following global registries on all hosts:
   • INSTPROF
   • PEER_DOMAIN

   For example

   installed_path/bin/db2greg -delvarrec service=GPFS_CLUSTER,variable=NAME,installpath=-

   installed_path/bin/db2greg -delvarrec service=PEER_DOMAIN,variable=NAME,installpath=-

   installed_path/bin/db2greg -delvarrec service=DEFAULT_INSTPROF,variable=DEFAULT,installpath=-

   installed_path/bin/db2greg -delvarrec service=INSTPROF,variable=<instance name>,installpath=-