Lock contention seen in Oracle multiple system readonly cache groups
Setup: We have a setup where 2 TimesTen 11.2.1.8.0 datastores on 2 different systems connect to an Oracle 11g database.
Both datastores have several readonly cache groups caching the same Oracle table.
In cases where 2 TimesTen datastores connect to one Oracle database, we see a wait event on the Oracle side called enq: UL – contention.
In Oracle database, we see continuous contention between TimesTen sessions from the two sources.
The TimesTen sessions are blocking each other continuously.
These sessions are among the top resource consuming sessions in the database.
We see exclusive locks on a table called TT_05_DB_PARAMS.
We don’t have this issue in a setup where one TimesTen source connects to one Oracle database.
Is it possible to avoid this contention at the Oracle level (when 2 TimesTen datastores connect to one Oracle database)?
We want to make sure that this wait event does not impact the performance of other processes in Oracle.
TimesTen Release 11.2.1.8.0 (64 bit Linux/x86_64)
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production