Re: sysviews corrupted

Yes we did ascii unloads and the database is completely on the disk we changed. And still I am reluctant to believe in a hardware problem : I only and always have problems with the sysviews table, each time again. For a hardware problem i would agree that such a problem could occur, but not always on the same table in the same database ( i have 3 databases in that chunk, but only 1 of them is a big one, and it is always on that one that it is happening ). Those 3 databases were reloaded in a different order each time, so it would be hard to believe (i think) in a hardware cause. convince me please, i need to make a statement in this by monday. Thanks. Guy BALLENTINE.MIKE@heb.com wrote: > > 2 questions: > > 1. You said you unloaded and reloaded all your data. Did you > do ascii unloads? If you used tbtape or tbunload, then > the corruption is still there. > > 2. You said in your original post that you replaced the disk. > Is your entire database on the disk you replaced? I had > corruption in an instance once that did not go away with > reinitializing. I ran ascii unloads, reinitialized > the instance, recreated the database, and reloaded. But > the corruption was still there. Only after I used dd to > copy /dev/zero over the offending chunk and then reinitialized, > did the problem go away. However, if you replaced the disk(s) > that the database uses, this is probably not the problem. > > Barring that, I would agree that it sounds like a hardware > problem. > > Good luck. > > Mike Ballentine > --------------------------( Forwarded letter 1 follows )--------------------- > Date: Friday, 26 September 1997 4:21am > From: germonpre_guy/dis@bekaert.com > Subject: Re: sysviews corrupted > Date: Fri, 26 Sep 1997 03:08:04 -0500 > Reply-To: Guy Germonpre <germonpre_guy/dis@bekaert.com> > Organization: Delaware Industrial Solutions > Sender: informix-list-owner@rmy.emory.edu > To: informix-list@rmy.emory.edu > X-Informix-List-Id: <news.43316> > > Art S. Kagel wrote: > > > > Guy Germonpre wrote: > > > > > > Hi, > > > > > > I have an Informix Online 4.10.UD2 on SCO 3.2v2.1. this thing worked for > > > a LONG time, but now i get errors "ptalloc bad page ..", and my sysviews > > > table is gone. > > > I unloaded all data, did a tbinit -i, and reloaded everything. > > > 2 days later the same problem. > > > I switched the disk. > > > 5 days later the same problem. > > > > > > This can't be bad luck. What is going wrong ? > > > > Do you use cooked files? Do you shut down the Informix Engine before > > shutting down or rebooting the machines either manually or through RC > > scripts? Have you had a crash that caused the machine to go down > > without executing normal shutdown? These are all possible causes of > > corruption. Others include a controller or power supply that is going > > bad and has become flaky. > > > > Art S. Kagel > > No we (my customer) does not use cooked files, the engine nor the > machine itself was not shut down prior to the error, no crash, just > suddenly a corrupted engine 3 or 4 times in a row. > The customer told me he had the error while creating temporary tables > via sequelink. There is anough disk space and database space so it is > not caused by shortage of disk. > > Do you really think it can be caused by HARDWARE ? > > Guy