Re: locking

In <01bd4bfd$5f328f40$6df3e7cc@service.beijing.informix.com> "Birdman" <fanjun@writeme.com> writes: >ANTS <sui@naseej.com> wrote in article <6e1dbo$bpd$1@news.xmission.com>... >> We have a 4GL program whenever it runs it will lock several tables. The >> problem is that the lock will remain long enough more than the deadlock >> timeout parameter. Why? and the other ppl's work will be locked as well. AFAIK, the DEADLOCK TIMEOUT parameter is only for distributed transactions. Is this the case with your problem? If not, then one application or the other will receive an error if a deadlock occurs. I suspect in your instance that the programs are not experiencing deadlock problems, but some other cause, like the transactions are not being ended. >> >> The isolation level is set to repeated read, is that appropriate? That's pretty application specific, both in terms of the program design, and the non-query load on the engine. If you don't have performance problems repeatable reads, then by all means keep it. -- Bryan Tonnet batonnet@phase4.com.au