> In other words, it's an artifact of the way Bucardo tries to > be safe with your data and can be ignored if it only happens > rarely.
There are 500 to 1000 errors each day. All those errors concern to INSERT INTO bucardo.bucardo_track (txntime,targetdb,tablename) SELECT DISTINCT txntime, 'node2'::text, 17640::oid FROM bucardo.bucardo_delta d WHERE d.tablename = 17640::oid AND NOT EXISTS ( SELECT 1 FROM bucardo.bucardo_track t WHERE d.txntime = t.txntime AND t.targetdb = 'node2'::text AND t.tablename = 17640::oid); with different tables OIDs > Those are the new serialization errors. Yes, it is new type of serialization error, but I wonder if something can be changed into bucardo code... 1) Is it a good idea to upgrade bucardo to new 4.99.x ? 2) Is it a good idea to add everything into only one sync.... ? as far I understand the basis of this bucardo error, it is a competition between concurrent transactions of bucardo itself. 3) Is it a good idea to rewrite this statement "INSERT INTO bucardo.bucardo_track ..." to locks less rows ? 4) Is it a good idea to reexecute this statement "INSERT INTO bucardo.bucardo_track ..." if such error occur ? 5) ...? What else, can someone suggests to do, to get rid of this type of errors... ? Before upgrading master node to 9.1.3, there were no errors. /Adam _______________________________________________ Bucardo-general mailing list [email protected] https://mail.endcrypt.com/mailman/listinfo/bucardo-general
