Re: 7.23 UC4 vs. 7.30 UC3

In article <6t71rs$nd5$1@nnrp1.dejanews.com>, icc@injersey.com writes > >> We are having problems with wrong passwords from setnet (I-net on PC's) >> causes the CPU vp to enter an infinite loop. >> >> Also some queries (probably only those with order by-clauses, but not >> all of them) seem to choose the wrong query path, regardless of the cost. >> This seems like a bug, but it is not verified by Informix yet. >> >> Lastly we have had 2 panics in oninit in 2 weeks. Nothing verified there, >> either. >> >> This is IDS 7.30.UC3 on HP-UX 10.20 with lots of patches. >> -- >> 'yvind Gjerstad Systems dept Tollpost-Globe AS N-6301 'ndalsnes/Norway >> E-mail: ogj@it.tollpost.no Phone: +47 7122 6663 Fax: +47 7122 6694 >> > >I've talked to an Informix software engineer today, who has been assigned to >our problem since we've become a squeeky wheel. She indicated that in 7.30 if >there is a thread problem then the thread will be pulled down and *not* the >engine. Also, she indicates that more detailed diagnostic information will be >available. > >QUESTION: does this sound bogus? We have had exactly the same information from Informix, following lots of Assert Fails with 7.23 on a system with a network known to be flaky in places. At first Informix thought we were killing front-end processes, but putting a log on the kill script ran that one out of the window. The site has had 7.30 for nearly a week with one problem - they had a strange block size on the tape devices and the first attempt at backing up the logical logs hung. However the block size was set to 16 and things have been OK since. <snipped> -- Surfer! http://www.nevis-vieww.demon.co.uk http://www.nevis-vieww.demon.co.uk/flash Hopeful anti-spam: alter double 'w' to single 'w' to view site & send Email.