Upgrade to 17.1 fails, often on configuring EEM

Document ID : KB000094254
Last Modified Date : 22/05/2018
Show Technical Document Details
Issue:
Upgrading Service Catalog to 17.1 fails.
This usually happens at the step where it connects to EEM but can also occur when configuring the MDB or somewhere else.
The install log will show the task task.configuring_eem in a failed state or task.configuring_mdb failed.
There is usually a memory error in the install log:
Exception in thread "main" java.lang.OutOfMemoryError: GC overhead limit exceeded
 
Environment:
Any upgrade to17.1 where the MDB contains large tables 
Cause:
This is caused by limitations in the JDBC driver updating the usm_lastid for tables which contain millions of rows.
This could be the usm_system_alert, usm_subscription_detail, or other tables.
This fails in the 'updateLastIds' method of commondb.java within common.core.jar

To see if you have large tables which will cause the issue, you can run:
select * from usm_lastid order by id desc;
 
Resolution:
Please contact support to get the  patches  T6D9365.CAZ  and  T6D9366.CAZ

SM 17.1 installer tried to  upgrade catalog  but  encountered the failure during upgrade which was caused by the reason mentioned here ,  then please do the following :
   1.apply T6D9365.CAZ
   2.re-invoke SM 17.1 installer to finish the upgrade to 17.1
   3.after upgrade finishes ,  apply T6D9366.CAZ    ( Note :  when 17.1 Rollup 1 patch is  available ,  apply R17.1 rollup1 patch  here instead of T6D9366.CAZ)


Note :  you need to apply T6D9365 to the failed upgrade box and then run the upgrade again as mentioned above .    The T6D9366 fix ( or 17.1 RU1 patch , it  will be available in June 2018 ) is only for after the upgrading successfully finishes .  

 
Additional Information:
To further verify that this solved the problem, you may want to check the status of your install and upgrade tasks by running:
select * from AL_CDB_COMPONENTINSTALLSTATE;