execution_engine_profiles

Patrick_T_Hildr

I'm still on V5.0 and I'm profiling a query that just runs and never returns or does anything after a few days at which point I have to kill it. Is there anything in particular i should be looking for in the results that show up in this table that will show what if anything its doing? Are there any other tables i should be looking at that will give m more info as to what vertica is trying to do with the query? It is several steps into a batch script that is using local temp tables along the way. To add some scale in for perspective one of the temp tables has about 200MM rows that is being LEFT JOINed to a few other multi-hundred million row tables in a subquery and inserting that into a staging table. I have been slowly replacing temp tables with perm tables along the way but it seems like this should be doable. Any suggestions here or should i open a support ticket? Thanks Patrick

Daniel_Leybovic

On each node, no resources are reserved or held while the query is in the queue. However, multi-node queries queued on some nodes will hold resources on the other nodes. HP Vertica makes every effort to avoid deadlocks in this situation.

try to monitor resource pools or system that you didn't get a dead lock.