Comment 2 for bug 1378127

From $MYSQROOT/logs/mvqr.log we see about 94 occurrences of this error on the test machine over the last 2 days. The instabce recorded here resulted in a table not error from Trafodion implying that we were scanning the OBJECTS table or its index. This table has about 1000 rows in the test machine and therefore it seems unlikely that any of the workarounds suggested above could actually help. The regionserver is showing no signs of this error in its logs. The 94 errors occurred in 2 waves, the first from 10/05 15:20 - 15:45, and the second from 10/05 22: 46 to 10/06 00:26. These could correspond simply periods when the test was run or more significantly when the system was under some kind of stress.

Next time tests are run it will help if we had a record of every time the error was raised, so that we know which table was being accessed. Also we can run with the setting log4j.category.HBase=DEBUG in $MY_SQROOT/logs/log4cpp.mvqr.config, to collect more debug information