Re: OnLine 5.00 database corruption

In article <OLIVER.93Nov19180330@test.infix.de> (Oliver Okrongli) writes: >We just experienced inconsistencies in OnLine 5.00 databases on two >different ports. It started with log messages "bfcheck: bad page ..." It IS a serious situation. You have a destroyed database. Your message stated that two different ports exhibited the same problem. Were both problems on the same machine? Were both ports run against the same disk drive and/or database? If your page addresses in the error text you posted are correct, you have a very large database (in excess of 32 chunks). Is this true? Does the error occur on the same chunk every time? We run very large databases (16-40GB) with >30 chunks. When we have seen corruption of this nature, it has almost always been hardware instability or incorrect configuration on our part. Are your chunks on RAW (character) devices? Is there any possibility of chunk overlap? (Check your disk slices!) Is there any possibility of two separate instances sharing the same space? Is there any possiblity of swap or file systems overlapping the Informix raw chunks? Did you have any crashes between the time the database was known good and the time it was discovered bad? Please let us know here if you or Informix discover the problem. Jim Burks jburks@promus.com Database Administrator