F
frank.meulblok
Guest
[quote user="Rob Fitzpatrick"] Happy to help. Here is some more info on client statement caching from the Progress Revolution 2011 conference: "What's That User Doing? The Progress OpenEdge Client Request Statement Cache" http://download.psdn.com/media/revolution_2011/oe11/OE1111.wmv [/quote] Statement caching isn't a great tool to track transaction scoping issues. You'll only get the point of the last database statement at the time you query it, which may or may not have anything to do with the statement that initiated or activated the transaction. Consider this fragment: DO TRANSACTION: FIND FIRST customer EXCLUSIVE-LOCK. customer.comments = customer.comments + " blah blah blah". PAUSE 15. FOR EACH salesrep NO-LOCK: END. PAUSE MESSAGE "check statement cache". END. If you check the cache when you're prompted to, you get the reference to the "FOR EACH salesrep". Also consider switching on client logging with the 4GLTrans entry type active. Keep logging level at 2 (basic), that will log only start and end of transactions, not substransactions. Keeps the logs as small as possible while still giving you what you need. If you spot a client with a long-running transaction, check their log for the last "BEGIN TRANS [code reference]" entry and tell the dev. team to start investigating from there.
Continue reading...
Continue reading...