[CONNECTORS-282] We see a lot of unique constraint collisions trying to insert into the ingeststatus table

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: ManifoldCF 0.4
• Fix Version/s: ManifoldCF 0.4
• Component/s: Framework agents process
• Assignee: Karl Wright
• Reporter: Karl Wright
• Created: 25/Oct/11 07:07
• Updated: 25/Oct/11 11:23
• Resolved: 25/Oct/11 11:23

Description

Postgresql logs the following over and over:

2011-10-24 23:48:56 EDT STATEMENT: INSERT INTO ingeststatus (id,changecount,dockey,lastversion,firstingest,connectionname,authorityname,urihash,lastoutputversion,lastingest,docuri) VALUES ($1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11)

This is clearly part of a failback sequence designed to prevent duplicates, but we seem to be erring by trying the least probable thing first. This no doubt impacts performance as well.