When replicating rows containing compressed partition blobs in
an Enterprise Replication environment an Assert Failure may
occur indicating memory corruption. An example AF is:
Pool 'CDRGeval1' (0x4665e040)
Bad free block 0x466aa458
Assert Failed: Memory free block header corruption detected in
mt_shm_malloc_segid 3
IBM Informix Dynamic Server Version 12.10.FC14
Who: Session(48, informix@server1, 0, 0x44f3ebe8)
Thread(75, CDRGeval1, 44efd6e8, 1)
File: mtshpool.c Line: 1834
Results: Unable to repair pool free list
Action: Please notify IBM Informix Technical Support.
stack trace for pid 29128 written to /dumpdir/server1/af.433b9ad
See Also: /dumpdir/server1/af.433b9ad
The contents of the AF file include:
Found during mt_shm_malloc_segid 3
Pool 'CDRGeval1' (0x4665e040)
Bad free block 0x466aa458
blk-64
00000000466aa418: 764f3854 47366179 6a463834 51423673 vO8TG6ay
jF84QB6s
00000000466aa428: 78545638 61446d35 4c717848 4a474842 xTV8aDm5
LqxHJGHB
00000000466aa438: 61745142 63663142 43202058 73203167 atQBcf1B
C Xs 1g
00000000466aa448: 69467533 75397a6e 54635657 744c4239 iFu3u9zn
TcVWtLB9
blk+64
00000000466aa458: 6d6a4f69 5320697a 6f344452 65723878 mjOiS iz
o4DRer8x
00000000466aa468: 6777755a 7377704a 4e484e20 78566972 gwuZswpJ
NHN xVir
00000000466aa478: 707a7362 38425174 7162674c 34714644 pzsb8BQt
qbgL4qFD
00000000466aa488: 38574945 705a3745 4d764d45 71716472 8WIEpZ7E
MvMEqqdr
Stack for thread: 75 CDRGeval1
afstack
afhandler
affail_interface
mt_shm_malloc_segid
mt_malloc
old_do_rhpstart
grouperEvalThread
cdrTrampolineThread
th_init_initgls
startup
The corrupted memory block headers may include evidence of the
uncompressed partition blob data.
Other types of Assert Failure are possible including those that
cause a fatal exception and crash the server. Example thread
stacks shown in such AFs include:
Stack for thread: 75 CDRGeval1
afstack
mt_ex_throw_sig
afsig_handler
fmclose
cdr_db_closetab
rqmCloseHandleInt
grouperEvalThread
cdrTrampolineThread
th_init_initgls
startup
Stack for thread: 269 sqlexec
afstack
afhandler
afcrash_interface
mt_free_mutex
destroy_session
sqsetconerr
_iread
_igetint
sqmain
On restart of the server it is also possible that further Assert
Failures may be raised by a CDRGeval thread, for example:
Assert Failed: No Exception Handler
IBM Informix Dynamic Server Version 12.10.FC14
Who: Session(34, informix@server1, 0, 0x44f3d748)
Thread(61, CDRGeval0, 44efe8a8, 1)
File: mtex.c Line: 508
Results: Exception Caught. Type: MT_EX_OS, Context: mem
Action: Please notify IBM Informix Technical Support.
See Also: /dumpdir/server1/af.425ba6f
Thread ID 61 will NOT be suspended because
it is deemed too critical to the server.
See Also: /dumpdir/server1/af.425ba6f
Stack for thread: 61 CDRGeval0
afstack
mt_ex_throw_sig
afsig_handler
ucvdrop
ucvreset
flfree
opfree
rsclose
cdrExpandRow
grouperCopy
grouperEvalThread
cdrTrampolineThread
th_init_initgls
startup
Recovery from this situation requires the server to be restarted
without Enterprise Replication and then the associated replicate
is made inactive. It may then be dropped and recreated followed
by a data resync activity to ensure all necessary data has been
replicated to the target table. |
