Has OnLine problems with Dell machines?

Hello, Informixer's. I hope somebody can help with this: I have two instances of On-line 5.03 UC1 running on a Dell Omniplex 560 with SCO UNIX 3.2.4.2. Each has its own raw device (created with divvy), and various dbspaces with appropiate offsets. One of the instances works finely, but the other produces an error when I try to unload data from a table (it has 75000 rows). The error reported is something about "cannot fetch next row", but the crazy thing generates the following log: 09:09:11 bfcheck: bad page: pg_stamp 18d14 != PG_STAMP2 18d62, userp = 80401b44, pid = 1863, uid = 200 09:09:11 buffer header: 09:09:11 80445a94: 00000000 00000000 f42c4480 f42c4480 ........ .,D..,D. 09:09:11 80445aa4: d4384480 10044080 02000000 e10c0000 .8D...@. ........ 09:09:11 80445ab4: 6f161000 00504980 00000000 00000000 o....PI. ........ 09:09:11 80445ac4: 01000000 54674480 00010000 ....TgD. .... 09:09:11 page header: 09:09:11 80495000: 6f161000 148d0100 48000100 d8060400 o....... H....... 09:09:11 80495010: 00000000 00000000 ........ 09:09:11 slot table and stamp: 09:09:11 804957fc: 628d0100 b... 09:09:11 -- Fail Consistency Check -- btpnext:bad page -- pid=1863 user=200us=80401b44 I've tried to rebuild the table, the database, the chunk, the dbspace an the whole On-line (including re-installing with another version), and the problem appears exactly the same all the times. And, worst of all, the diagnostics software doesn't report any failures on the hard disk or on the controller. But that's not all! After the last error, I cannot start tbmonitor! I get "memory fault - core dumped" with both Online instances. I have exactly the same trouble on a twin machine, with identical configuration. But all works fine if I try on a Dell 4066/XE or on a Compaq System Pro/LT. If anyone has an idea, or knows about some incompatibility between Dell Pentiums and OnLine, please let me know. Thanks in advance. ----------------------------- Juan Manuel De Villeros Arias ejda142@delta.eafit.edu.co