Re: Strange Log behavior

In article <93ngo5$a6f$1@news.panix.com>, Michael Hoffman <mrh@panix.com> wrote: > Hi all, > I'm noticing something VERY strange on our test system. This is a > small SUN system (1 CPU), running SUN OS 5.7 and Informix 7.2.3 (I know, it's > outdated! We'll upgrade to 9.2 very soon). > > We use the test system solely for benchmarking right now. I have a > program which updates 91122 records in one table. When that finishes, I then > fork off 9 process which do more updates on the same table, simultaneously. > This test is definitely not efficient, but it is a small simulation of what Is the 91122 update all in one transaction?(I am guessing it is) If so, informix can not free up any logical logs which contain records of this transaction until the transaction complete. If that is the issues, break up the updates into multiple transactions... If my guess is wrong posting the onstat -l and onstat -x output from when the logs are full might help narrow down what is going on. Hope this helps, Will <SNIP> > Sent via Deja.com http://www.deja.com/