Re: bfcheck: bad page: pg_stamp ... != PG_STAMP2 ...

> >Hi, >Could you help to find solution for our problem or make reference on known >bug: > >1) 1-2 times per week we are having this kind of error: > >18:25:43 bfcheck: bad page: pg_stamp 47ab9f02 != PG_STAMP2 9f025c1b>buffer header 0de86dfc: 0de86e64 0de86d94 00000000 00000049 ..nd..m. >........I 0de86e0c: 0000013a 00000000 00000000 0007c17c ...:.... .......| >0de86e1c: 07d10001 00006266 5b313830 3332355d ......bf [180325] >0de86e2c: 00000000 00000000 0c6dbfb4 0c6dbfb4 ........ .m...m.. >0de86e3c: 0a00e55c 0d9dcad4 00020000 00720000 ...\\.... .....r.. >0de86e4c: 05c48deb 242c9000 00000000 00000000 ....$,.. ........ >0de86e5c: 00010010 69377528 ....i7u > page 242c9000: 05c48deb 47ab9f02 I could not see the entire page. Normally on this level, the af.xxx file will contain 1) the page that it had read followed by 2) a physical re-read of the same page from disk. I would be interested in seeing what that contained. Was it the same page? What does the page on disk actually look like? This can be gotten from oncheck -pP 0x5c 0x48deb. Is there a case for this problem? >002e0050 06a003c2 ....G... ...P.... 242c9010: 000428c0 000317de 30313030 >30303030 ..(..... 01000000 242c9020: 31303938 32343200 027bfa30 31303030 > 1098242. .{.01000 ' 242c97e0: 0013067a 0013003e 00130667 0013002b >....z...> ...g...+ 242c97f0: 00130524 00130018 001347ab 9f025c1b ...$.... >...G...\\. chunk info 278afce0: 005c005d 00000008 000f9ff8 00004068 >..\\.].... ......@h 278afcf0: 00030000 00400026 2f696e66 6f726d69 .....@.& >/informi ' 5c15b7f0: 002b0013 05240013 00180013 47ab9f02 .+...$.. >.....G... 18:25:43 18:25:43 Assert Failed: Page Check Error in >btsearch:bad btree node 18:25:43 Who: Session(17874, sapr3@ohFDDI, >22348, 663461848) Thread(21496, sqlexec, 2c137918, 7) 18:25:43 Results: >Possible inconsistencies in index 'prd:"sapr3".aufm-Key#2' 18:25:43 >Action: Run 'oncheck -cI 3145791' 18:25:43 See Also: /crash/af.53f83bf7, >shmem.53f83bf7.0, gcore.53f83bf7.0 18:25:43 Stack for thread: 21496 >sqlexec ' > >2) Recommended action 'oncheck' did not show any errors. And the same >tables never had this problem twice. > >It could be different table and different message, but always having text >'bfcheck: bad page: pg_stamp ' != PG_STAMP2 '' > >We have INFORMIX-OnLine Version 7.23.UC6 on Solaris (SunOS yaz 5.5.1 >Generic_103640-12 sun4u sparc SUNW,Ultra-Enterprise) > > >Thank you > >-Joseph Shkolnik, >shkolnij@hotmail.com > >-------------------==== Posted via Deja N Madison Pruet