1 Reply Latest reply on Sep 5, 2012 4:51 PM by Mike Albertone

    Issue with accumulating logs on master DS in a replicated env

    Mike Albertone

      We have a production system running TimesTen Release on SunONS 5.10. We use it in a replicated environment and we got an accumulation of logs on one of the DS. THey are symmetrically defined as both master and subscribers (for each other) and one of them is showing an accumulation of logs due to the following entry in SYS.TRANSACTION_LOG_API:

      Command> select * from SYS.TRANSACTION_LOG_API;
      < DBCNbookmark , 891082, 5601408, 891082, 5601344, 23054, 00 >

      Command> call ttLogHolds();
      < 891082, 5601344, XLA , DBCNbookmark > <<<<<<<<<<<<<<<
      < 891457, 6397872, Checkpoint , tt41data.ds1 >
      < 891461, 3148680, Checkpoint , tt41data.ds0 >
      < 891462, 3755752, Replication , RMNPN03-TTREP:TT41DATA >

      Currently the logs are kept starting from tt41data.log891082 to tt41data.log891423 and growing.

      I can see that the PID 23054 has gone (system got an issue yesterday night) so that don't know how to cleanup/force that transaction in order to let the rest of the transactions being pushed to the other side and logs being cleaned.

      Is there any procedure for that?