Update from 10.00UC1 to 10.00UC4 poops out

I've had an IDS instance on an x86/LINUX box since 7.x something and it has always been *ROCK* stable, upgrades to 9.21 to 9.4x to 10.00UC1 have always gone off without a hitch. Now I've got a minor upgrade that has bit it twice, from 10.00UC1 to 10.00UC4. Reverting back to 10.00UC1 and reloading the databases and all is well. Anyone have any ideas: Sequence ---- 1. Exported the two databases with dbexport. 2. Ran a series of onchecks to see that there wasn't any problems. oncheck -q -cr oncheck -q cR oncheck -q ce oncheck -q cc miecr oncheck -q cc cms oncheck -q -x -ci miecr oncheck -q -x -ci cms oncheck -q -x -cI miecr oncheck -q -x -cI cms oncheck -q -cD miecr oncheck -q -cD cms .. these all completed without complaint. 3. Took the server offline. 4. Ran ""./ids_install" from C880JNA.tar 5. Started server via onmonitor 6. "onstat -d" shows all chunks online, queries seem fine, tested with applications. Dropped and recreated some indexes, etc... 7. Start a test with a long running 4GL application, it runs for a very long time (approx four hours) and then dies with - duperr program fails with - Program stopped at "flowcntrl.4gl", line number 232. SQL statement error number -243. Could not position within a table (informix.xrefr). SYSTEM error number -105. ISAM error: bad isam file format. The output of "onstat -m" looks like - 09:58:41 IBM Informix Dynamic Server Version 10.00.UC4 09:58:41 Who: Session(109, intranet@KOHOCTON.morrison.iserv.net, 6814, 0x4b7a04a0) Thread(2895, sqlexec, 4b77c06c, 1) File: rsdebug.c Line: 1092 09:58:41 Results: Possible inconsistencies in '11121E1700 e:"11121U316071 \\uffff\\uffff".11122 $\\uffff' 09:58:41 Action: Run 'oncheck -cDI 11121E1700 e:"11121U316071 \\uffff \\uffff".11122 $\\uffff' 09:58:41 stack trace for pid 2232 written to /opt/informix/dump/af.f37fc10 09:58:41 See Also: /opt/informix/dump/af.f37fc10, shmem.f37fc10.0 09:58:44 Unable to create output file '/opt/informix/dump/shmem.f37fc10.0' errno = 28 09:58:44 Page Check Error in btcurrent:bad current page 09:58:45 Assert Failed: Page Check Error in btcurrent:bad current page 09:58:45 IBM Informix Dynamic Server Version 10.00.UC4 09:58:45 Who: Session(109, intranet@KOHOCTON.morrison.iserv.net, 6814, 0x4b7a04a0) Thread(2895, sqlexec, 4b77c06c, 1) File: rsdebug.c Line: 1066 09:58:45 Results: Possible inconsistencies in index '11121E1700 e:"11121U316071 \\uffff\\uffff".11122 $\\uffffKey#1' 09:58:45 Action: Run 'oncheck -cI 11121E1700 e:"11121U316071 \\uffff \\uffff".11122 $\\uffffKey#1' 09:58:45 stack trace for pid 2232 written to /opt/informix/dump/af.f37fc10 09:58:45 See Also: /opt/informix/dump/af.f37fc10 09:58:49 Page Check Error in btcurrent:bad current page Eh? Dump directory unfortunately fills up (subsequently changed to somewhere with additional disk space). "onstat -d" now shows that some chunks are offline - address chunk/dbs offset size free bpages flags pathname 4b689948 1 1 0 976885 856593 PO-B /opt/informix/dev/root_chunk 4b689ad0 1 1 0 976885 0 MO-B /opt/informix/dev/root_chunk_mirror 4b79bd38 2 2 1 8885090 6680557 PO-B /opt/informix/dev/datadbs_chunk 4c1ed948 2 2 1 8885090 0 MO-B /opt/informix/dev/datadbs_chunk_mirror 4c1ed018 3 3 0 8886776 ~8886776 8886776 POBB /opt/informix/dev/blobspace0 4c1edad0 3 3 0 8886776 0 MOBB /opt/informix/dev/blobspace0_mirror 4c1ed1a0 4 4 0 8886776 0 PD-B /opt/informix/dev/indexdbs 4c1edc58 4 4 0 8886776 8326382 MO-B /opt/informix/dev/indexdbs_mirror 4c1ed328 5 5 0 1000000 0 PD-B /opt/informix/dev/tmpdbs_chunk2 4c1ed4b0 6 3 0 8886776 ~8886776 8886776 POBB /opt/informix/dev/blobspace1 4c1edde0 6 3 0 8886776 0 MDBB /opt/informix/dev/blobspace1_mirror 4c1ed638 7 5 0 1000000 0 PD-B /opt/informix/dev/tmpdbs_chunk3 4c1ed7c0 8 5 0 100000 0 PD-B /opt/informix/dev/tmpdbs_chunk 8 active, 32766 maximum Shutdown server, restored backup of /opt/informix, reinitialized first chunk of raw device, added back chunks and dbspaces, imported databases from the dbexport via dbimport. Done this twice now.