[ https://issues.apache.org/jira/browse/TRAFODION-1748?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Suresh Subbiah resolved TRAFODION-1748. --------------------------------------- Resolution: Fixed > Error 97 received with large upsert and select statements > --------------------------------------------------------- > > Key: TRAFODION-1748 > URL: https://issues.apache.org/jira/browse/TRAFODION-1748 > Project: Apache Trafodion > Issue Type: Bug > Components: dtm > Affects Versions: 1.3-incubating > Reporter: Sean Broeder > Assignee: Sean Broeder > Priority: Major > Fix For: 2.0-incubating > > > From Selva- > The script has just upserted 100000 rows and querying these 10000 rows > repeatedly. From the RS logs, it looks like memstore got flushed. Currently, > I have made the process to loop on getting the error 8606. > > This query involves ESPs. This error is coming from sqlci at the time of > commit. I assume sqlci must be looping. The looping ends after 3 minutes to > proceed further. You can also put sqlci into debug and set loopError=0 to > come out of the loop to proceed further. I also created a core file of sqlci > at ~/selva/core.44100. > > If the query is finished, you can do the following to reproduce this issue > > cd ~/selva/LSEG/master/stream > sqlci > log traf_stream_run.log ; > obey traf_stream_run.sql ; > log ; > ++++ > Looking at dtm tracing I can see the regions are throwing an > UnknownTransactionException at prepare time, which causes the TM to refresh > the RegionLocations and redrive the prepare messages. These again fail and > the transaction is aborted and this eventually percolates back to SQL as an > error 97. -- This message was sent by Atlassian JIRA (v7.6.3#76005)