Re: Failed Consistency Checks

In article <Crt3o1.7J3@alert.com>, ewh@alert.com (Ed Hobbs) says: > >Background: We're running OnLine 5.00, I-Star 5.00 and 4GL/SQL 4.10 > on RS/6000's. Our database is 12 GBytes in size composed > of over 350 tables. >09:18:07 bfcheck: bad page: pg_flags 0 != type 2, userp = 300056ac, pid = 78799, uid = 443 >09:18:07 buffer header: >09:18:07 30c5d7b8: 00000000 00000000 00000000 30bb2094 ........ ....0. . >09:18:07 30c5d7c8: 30bb2094 30c7e138 30c6fdf8 00030000 0. .0..8 0....... >09:18:07 30c5d7d8: 10170000 001000db 3240d000 00000000 ........ 2@...... >09:18:07 30c5d7e8: 00000000 00010000 30ccd92c 01000000 ........ 0..,.... >09:18:07 page header: >09:18:07 3240d000: 001000db 1e325049 00000000 00180fe4 .....2PI ........ >09:18:07 3240d010: 00000000 00000000 ........ >09:18:07 slot table and stamp: >09:18:07 3240dffc: 1e325049 .2PI >09:18:07 tblspace header error: >09:18:07 3085d218: 00000000 00000000 00000000 00000000 ........ ........ >09:18:07 3085d228: 00010000 010000ce 001000db 00000000 ........ ........ >09:18:07 3085d238: 00000000 00000000 00000000 00000000 ........ ........ >09:18:07 3085d248: 00000000 00000000 00000000 00000000 ........ ........ >09:18:07 3085d258: 00000000 00000000 00000000 00000000 ........ ........ >09:18:07 3085d268: 00000000 00000000 00000000 00000000 ........ ........ >09:18:07 3085d278: 00000000 00000000 01000049 0921000a ........ ...I.!.. >09:18:07 3085d288: 003e0003 00010001 2e0855f8 00000001 .>...... ..U..... >09:18:07 3085d298: 00000008 00000008 00000008 00000003 ........ ........ >09:18:07 3085d2a8: 00000001 ffffffff 00000037 00000001 ........ ...7.... >09:18:07 3085d2b8: 00000000 00101b55 00000008 00000000 .......U ........ >09:18:07 3085d2c8: 00000000 00000000 00000000 00000000 ........ ........ >09:18:07 3085d2d8: 00000000 00000000 00000000 00000000 ........ ........ >09:18:07 3085d2e8: 00000000 00000000 00000000 00000000 ........ ........ >09:18:07 3085d2f8: 00000008 00000000 00000000 00000000 ........ ........ >09:18:07 3085d308: 00000000 30aba4a8 1e32506f 00000000 ....0... .2Po.... >09:18:07 3085d318: 30b393e8 0... >09:18:07 -- WARNING: Fail Consistency Check -- pthdrpage:ptalloc:bad partn page -- pid=78799 user=443 us=300056ac >09:21:14 Checkpoint Completed >09:31:30 Checkpoint Completed >09:42:12 Checkpoint Completed >======================================================================> >We have been getting an increasing number of these errors in our turbo.log, >however tbcheck -pt, -cd and -ci do not indicate any problems. > >Informix Support recommends that we dbexport the entire database and reload it >but I estimate this will take 2-3 days. > >Does anyone know anything more about this error and maybe how to find and >correct the bad pages? Can you isolate it to a table, and a specific SQL statement? In the example you have, the first data number under "Page header:" is 001000db. Is that always the case? Is the number usually "nearby". Ask Informix tech support to have one of the Advanced Support engineers look at it. Sometimes, they can patch the error without having to reload. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~ Jim Burks jburks@promus.com ~ ~ Database Administrator / Systems Engineer ~ ~ The Promus Companies, Inc. Memphis, TN USA ~ ~ http://stargate.promus.com/public/jbb.html ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~