Tablespace, loosing my mind

Hi, Informix combattants, I think I have a (possibly two) Sun special problem with my Informix environment: SunOS 4.1.3 INFORMIX-4GL Version 4.10.UC2 INFORMIX-SQL Version 4.10.UC2 INFORMIX-OnLine Version 5.00.UC1 A few weeks ago I reported a serious problem: Every now an then my application would crash because of an error at a part of my program where I just had checked all possible sources of errors. To my amazement I figured out that the error was the fact that the tablespace was gone. I described my problem to you, fellow netters, and I received answers from three people I want to say "thank you" hereby. They are Bob Baskett <inf_bb@hermes1.sps.mot.com> Graeme Sargent <graeme@pyra.co.uk> Jack Parker <jparker@hpbs2561.boi.hp.com> Mr Sargent mainly flamed Mr Baskett for his solution and Mr Parker shared my grief about the fact that the tablespace vanishes in a queer manner. None of us, still, knows a remedy against the fact that the tablespace gets lost, all hints were about how to repair a database with a vanished tablespace. Well, I backuped more often, and reinstalled the whole thing everytimes it happened, I lost faith completely, but it worked for me (well, I am the developer, and I shake and tremble with fear when that part of software is to be installed at my client's machines) Friends, I don't weant to repair the database, I want to know how things like that happen! I don't do nasty things in my program like raw device access or patching tables... I write plain 4GL code. But: it is always the same table, one out of thirty, and always the same module that causes the problem, but posting the code would distract you attention from what really is going on: My dogmatic belief is that I _cannot_ destroy a database by only using SQL. I didn't give in too easily. What I tried: I broke down the size of the transactions, no help, I moved the Data Segment to and fro using fillers, I installed the whole beast on a File chunk (not a raw device). I found out that the foreach construction doesn't stop if the status is -400 (what a nasty bug!), but now I'm fed up. It would be of great help to me if someone out there tells me either "I had the same problem like you, the solution was ..." or even "I have the same environment like you, but the problem did never show up here", in the latter case I'd swap my hardware for a change. There is another problem, maybe related to this: My Sun is heavy loaded when the tbinit daemon is running, even if noone actually works with informix, and from times to times all blocks and looks like freezed, then the load is going up to 3,4 or even higher (no other processes running! Well, you know what I mean, of course are there processes) The problem seems not to show when I start the window manager remotely, but thus I need two Suns for one developer's workplace, what a waste. Now I'm working at a terminal, it faster (rats), but the vanishing tablespace phenomenon ("VTP" we can call it) is still there. Hopeless, but not desparate, like so many times ;-) -svb. -- Sebastian v. Bomhard, +49-89-298934, +49-89-296629 <svb@muc.de> St.-Anna-Str. 13, D-80538 Muenchen (Lehel) >>>>>>>>>>> The early bird catches the EARLY worm. <<<<<<<<<<<<