Configuration has 3 source databases. The sp_cop has three captures and 3 readers but only one export queue . If running one capture at the time then everything runs.
event_log
Error 2011-10-17 14:27:00.113556 5117 1699599024 Reader killed due to SIGSEGV, core file = /global/plsplex01/apps/plsplex01/var/dump/core.7676 was not generated, pid = 7676 (from {source sid})
reader log
ord 2011-10-20 14:32:59.353920 18544 1206437536 rim_write(rmsg) mtranid=2 mlen=344 mpseq=483591000 rsid=a0325360.7.0
ord 2011-10-20 14:32:59.353936 18544 1206437536 Message written mid=1195136, subque id = 0 op_type= 6
ord 2011-10-20 14:32:59.353945 18544 1206437536 message seqno/offset = 3129/29181916 rowid=AAAENAAAGAAAAJ3AAB
ord 2011-10-20 14:32:59.353953 18544 1206437536 rim_write(rmsg) mtranid=2 mlen=344 mpseq=483591000 rsid=a0325360.7.0
ord 2011-10-20 14:32:59.353976 18544 1206437536 que_NODATA
core.out
Core was generated by `/~/.app-modules/sp_ordr -uINT10'.
Program terminated with signal 11, Segmentation fault.
#0 0x00002b1f3914f725 in bu_wt (que_seg_base=0x2b1f3ec9c000,
bp=0x2b1f3eca7570) at ../src/que/sp_que.c:7567
in ../src/que/sp_que.c
#0 0x00002b1f3914f725 in bu_wt (que_seg_base=0x2b1f3ec9c000,
bp=0x2b1f3eca7570) at ../src/que/sp_que.c:7567
#1 0x00002b1f39146b3e in sque_writecommit (QUE_p=0x2b1f3d53c4c0,
sque_p=0x2b1f3ec9c040, rw_index=0, chkpt=47413152057472)
at ../src/que/sp_que.c:3468
#2 0x00002b1f391468f9 in que_writecommit (qh=0x2b1f3ec9c040, chkpt=0)
at ../src/que/sp_que.c:3405
#3 0x00002b1f3915aae8 in rim_writecommit (chkpt=0)
at ../src/rim/rim_exec.c:491
#4 0x000000000040d79d in q_writecommit () at ../src/ordr/sp_ordr.c:1377
#5 0x000000000040db37 in checkpoint (wcache=0) at ../src/ordr/sp_ordr.c:1499
This SIGSEGV occurred during queue writecommit. The export queue has multiple sp_ordr processes as queue writers
Example:
Source:
qview -i
qview> qinit
qview> list
The following queues exist:
o.sid1+C
WRITER +PA+o.sid1+sp_ocap+o.sid1
READER +PR+o.sid1+sp_ordr+o.sid1{name}+X
WRITER +PR+o.sid1+sp_ordr+o.sid1
READER +PX+{name}+sp_xport+0xa032532b (xxx.xx.xx.xx)
WRITER +PR+o.sid2+sp_ordr+o.sid2
WRITER +PR+o.sid3+sp_ordr+o.sid3+o.sid2+C
WRITER +PA+o.sid2+sp_ocap+o.sid2
READER +PR+o.sid2+sp_ordr+o.sid2+o.sid3+C
WRITER +PA+o.sid3+sp_ocap+o.sid3
READER +PR+o.sid3i+sp_ordr+o.sid3
Multiple queue writers would cause queue corruption since they would be writing conflicting information to the queue and subqueue headers during writecommit.
Re-configure the config file with named export queues and activate the new config.
For more information and step by step instructions refer to Shareplex for Oracle Administrator’s Guide, chapter "Configuration for isolating data using named export queues".
