It's possible when trying to delete a replicate that the
CDRCparse thread can fail in the following way:
09:32:24 CDR Grouper: Grouper evaluation blocked
09:32:24 stack trace for pid 1871 written to
/opt/informix/tmp/af.41a3488
09:32:24 Assert Failed: No Exception Handler
09:32:24 IBM Informix Dynamic Server Version 11.70.FC8
09:32:24 Who: Session(23, informix@server, 0, 0x4b9853c0)
Thread(50, CDRCparse, 4b948708, 1)
File: mtex.c Line: 490
09:32:24 Results: Exception Caught. Type: MT_EX_OS, Context:
mem
09:32:24 Action: Please notify IBM Informix Technical Support.
09:32:24 See Also: /opt/informix/tmp/af.41a3488,
shmem.41a3488.0
09:32:24 Stack for thread: 50 CDRCparse
base: 0x0000000050b94000
len: 69632
pc: 0x0000000001317a03
tos: 0x0000000001c2a990
state: running
vp: 1
afstack
mt_ex_throw_sig
afsig_handler
itexecinit
closesdb
doreleasesdb
gcClearStmt
gcReplParseThread
cdrTrampolineThread
th_init_initgls
startup
This happened in the case where an immediately previously
submitted cdr define replicate command was submitted and failed
with a -208 error (this was caused by hitting the max number of
local replicate tables already being in use). That failure gets
a portion of the CDRCparse thread structures into a state that
when the cdr delete replicate command was executed, the
CDRCparse thread then SEGV's and assert fails. |
