alter from serial to int crashes server

Posted in 2006

DCP — Sun, 13 Aug 2006 20:52:46 -0400 — source: Usenet: comp.databases.informix

Why would this (converting doc_id from type serial to type integer): alter table docs modify doc_id integer not null constraint doc_nn1; cause this: 20:33:48 stack trace for pid 6469 written to /data/docs1/informix/af.d9dcc4eb 20:33:48 Assert Failed: No Exception Handler 20:33:48 IBM Informix Dynamic Server Version 9.40.UC8 20:33:48 Who: Session(119617, informix@server.local, 5978, 0x61b54ef8) Thread(120308, sqlexec, 61b377a0, 1) File: mtex.c Line: 435 20:33:48 Results: Exception Caught. Type: MT_EX_OS, Context: mem 20:33:48 Action: Please notify IBM Informix Technical Support. 20:33:48 stack trace for pid 6469 written to /data/docs1/informix/af.d9dcc4eb 20:33:48 See Also: /data/docs1/informix/af.d9dcc4eb 20:33:51 mtex.c, line 435, thread 120308, proc id 6469, No Exception Handler. 20:33:52 The Master Daemon Died 20:33:52 PANIC: Attempting to bring system down ??? This problem was one reason we went from 9.40.UC5 to 9.40.UC8. IBM support has had no ideas other that dropping all indices (did not help). We are running on RHES 3 (Taroon Update 7). -- DCP

↪ replying to DCP

Jonathan Leffler — 14 Aug 2006 00:31:07 -0700 — source: Usenet: comp.databases.informix

DCP wrote: > Why would this (converting doc_id from type serial to type integer): > > alter table docs modify doc_id integer not null constraint doc_nn1;> > cause this: > > 20:33:48 stack trace for pid 6469 written to > /data/docs1/informix/af.d9dcc4eb > 20:33:48 Assert Failed: No Exception Handler > 20:33:48 IBM Informix Dynamic Server Version 9.40.UC8 > 20:33:48 Who: Session(119617, informix@server.local, 5978, 0x61b54ef8) > Thread(120308, sqlexec, 61b377a0, 1) > File: mtex.c Line: 435 > 20:33:48 Results: Exception Caught. Type: MT_EX_OS, Context: mem > 20:33:48 Action: Please notify IBM Informix Technical Support. > 20:33:48 stack trace for pid 6469 written to > /data/docs1/informix/af.d9dcc4eb > 20:33:48 See Also: /data/docs1/informix/af.d9dcc4eb > 20:33:51 mtex.c, line 435, thread 120308, proc id 6469, No Exception > Handler. > 20:33:52 The Master Daemon Died > 20:33:52 PANIC: Attempting to bring system down > > ??? This problem was one reason we went from 9.40.UC5 to 9.40.UC8. IBM > support has had no ideas other that dropping all indices (did not help). > > We are running on RHES 3 (Taroon Update 7). There is no obvious good reason for that to cause trouble. You should have contacted IBM Informix Tech Support, as the message says - and as you did. If the problem is reliably reproducible, and you've done obvious things like run ON-Check on the table and its indexes, then you should get a better response - like a PMR - from Tech Support. If the first line of support have no ideas (which is probably fair), then it should be propagated up through the hierarchy. The only area that could give cause for concern is what RedHat did in Update 7 (and maybe updates 1-6 as well). I don't know sufficient about that. One thing to check carefully is what version of Linux (kernel version and glibc version) 9.40.UC8 was ported to, and how this compares with the versions on your system. If you are behind in any versions, worry a lot. If your versions are the same or newer, there is less cause for concern.

↪ replying to DCP

scottishpoet — 14 Aug 2006 09:16:00 -0700 — source: Usenet: comp.databases.informix

Do you have a PMR number from when this was reprted to IBM Support?

↪ replying to scottishpoet

scottishpoet — 15 Aug 2006 02:21:39 -0700 — source: Usenet: comp.databases.informix

scottishpoet wrote: > Do you have a PMR number from when this was reprted to IBM Support? if you drop distributions on the table UPDATE STATISTICS LOW FOR TABLE <table> DROP DISTRIBUTIONS and then run the ALTER TABLE command does this stop the crash from occuring??