When creating an RSS or HDR secondary server and rolling forward
through the logical logs multiple Assert Failures may occur if
the log records contain truncate operations relating to
temporary partitions. An example AF is:
10:40:07 Assert Failed: Page Check Error in
trunc_mappages_free: bad bitmap page
10:40:07 IBM Informix Dynamic Server Version 11.70.FC7XOB4
10:40:07 Who: Session(20, informix@server1, 0, 0x44feed68)
Thread(41, xchg_1.2, 44fb5170, 1)
File: rsdebug.c Line: 1075
10:40:07 Results: Possible inconsistencies in 'TBLSpace
2097222'
10:40:07 Action: Run 'oncheck -cD 2097222'
10:40:07 stack trace for pid 2100 written to /tmp/af.411fe76
10:40:07 See Also: /tmp/af.411fe76
10:40:08 Rollforward of log record failed. iserrno = 105
10:40:08 Log Record: log = 6, pos = 0x53f2b0, type =
OLDRSAM:TRUNCATE(119), trans = 25
with the contents of the AF file including:
10:40:07 bfcheck: bad page: pg_addr 0:0 != bp->bf_pagenum
2:1583
..
page
000000004456b000: 00000000 00000000 00000000 00000000 ........
........
000000004456b010: 00000000 00000000 00000000 00000000 ........
........
..
10:40:07 Stack for thread: 41 xchg_1.2
base: 0x000000004b1e8000
len: 69632
pc: 0x0000000001313673
tos: 0x000000004b1f7cd0
state: running
vp: 1
(oninit) afstack
(oninit) afhandler
(oninit) affail_interface
(oninit) bffail
(oninit) trunc_mappages_free
(oninit) truncpn
(oninit) dotruncate
(oninit) plogredo
(oninit) rlogm_redo
(oninit) next_recvr
(oninit) producer_thread
(oninit) startup
The section of the AF showing the page contents as read from
disk shows it to be blank suggesting that it was not on the
archive image.
Other Assert Fail messages may also appear in the online log
referencing the table name "_temptable"
This problem is associated with SQL SELECT statements using the
CONNECT BY syntax that were active at the time of the archive.
Once the secondary server has successfully synchronized with the
primary then running these statements does not cause a problem. |
