OpenSolaris

Printable Version Enter a New Search
Bug ID 6529966
Synopsis The database preparation does not work when the master server is restored from a different cluster
State 10-Fix Delivered (Fix available in build)
Category:Subcategory suncluster:ha-sps
Keywords
Responsible Engineer Detlef Ulherr
Reported Against
Duplicate Of
Introduced In
Commit to Fix 3.2u1_01
Fixed In 3.2u1_01
Release Fixed 3.2u1_fcs(3.2u1_01)
Related Bugs
Submit Date 1-March-2007
Last Update Date 30-May-2007
Description
Restoring a master server from abackup in a different cluster, does not restore the pair monitoring table / monitoring user in a consistant way. The necessary database preparation fails, because the user and the table still exists. However the old monitoring user does not own the montoring table in reality. Any resource enable without successful repreparation will fail. 

Becaused of this problem, the database preparation needs to drop the monitoring user and the monitoring table before recreating them.
Work Around
1. Restore the master server.
2. Start the master server manually.
3. Cd to your server bin directory .
4. Invoke a psql with ./roxdbcmd psql rox .
5. Drop the user with drop user sc_test;
6. Drop the table with drop table sc_test;
7. Teminate the psql.
8. Prepare the  database again. If the drop table and drop user were
successful the prepare would succeed.
9. (optional) determine the probe command with scrgadm -pvv orclresource show -v
10. (optional) you may try the probe command of the master servers ha resource with  ksh it should return 0 now.
11. shut down the master server
12. enable the master servers resource.
Comments
N/A