IDS Forum

Bitmap is wrong for partp in sbspace

Posted By: Peter Lempochner
Date: Wednesday, 7 August 2013, at 9:51 a.m.

Hello,

we have a problem with inserting/updating SBLOBs in one of four sbspace.
There are many AFs in online log (all AFs describes the same issue):

==========================
15:55:10 IBM Informix Dynamic Server Version 11.70.FC3 Software Serial
Number AAA#B000000

15:55:10 Assert Warning: Bitmap appears to be corrupt
15:55:10 Who: Session(2842, ezu@::ffff:10.10.7.122, -1, 70000030e62dff0)

Thread(3414, sqlexec, 7000002f543e1b0, 1)

File: rsbitmap.c Line: 2515
15:55:10 Raw hex dump of stack located in /usr/idb/tmp/af.113e003d.rawstk
15:55:10 Stack for thread: 3414 sqlexec

base: 0x070000037ae8b000

len: 69632

pc: 0x000000010008c2ec

tos: 0x070000037ae98ce0
state: running

vp: 1

0x000000010008c098 (oninit)afstack
0x000000010008b63c (oninit)afhandler
0x000000010008b9ec (oninit)afwarn_interface
0x000000010036ac2c (oninit)allocvrow
0x00000001005a3134 (oninit)genpg_insertx
0x00000001005a22a4 (oninit)genpg_insert
0x00000001008acb30 (oninit)sba_loheader_new
0x00000001008ad434 (oninit)sba_loheader_create_wlock
0x00000001008a9834 (oninit)sb_create
0x000000010089e9e0 (oninit)sbx_create
0x00000001008973d0 (oninit)sq_lo_create
0x0000000100e1fed4 (oninit)udrlm_clang_execute_internal
0x0000000100e1fa08 (oninit)udrlm_clang_execute
0x0000000101397df0 (oninit)udrlm_exec_routine
0x000000010057302c (oninit)udr_execute
0x0000000100231450 (oninit)sq_exfproutine
0x0000000100431b70 (oninit)sqmain
0x0000000100a42a08 (oninit)listen_verify
0x0000000100a41abc (oninit)spawn_thread
0x00000001013fbf40 (oninit)startup

15:55:10 See Also: /usr/idb/tmp/af.113e003d

---------------------------------
Begin System Alarm Program Output
---------------------------------

Assertion Failure Type: Warning
Host Name: xezur
Database Server Name: ses
Time of failure: Tue Aug 6 15:55:10 CEST 2013
AF file: /usr/idb/tmp/af.113e003d
Shared memory file: None
System Blocking: OFF

-------------------------------
End System Alarm Program Output
-------------------------------

15:55:11 sh /usr/idb/etc/evidence.sh 1 0 /usr/idb/tmp/af.113e003d 2842
0x7000002f543e1b0 3414 0x7000002e0138d40 1025 0 0 0 0
15:55:11
------------------ End of assertion failure 0 -----------------

15:58:04 Bitmap is wrong for partp 0xdc3e94e0, partnum 19922949<<< This
is our default sbspace (some schemas have PUT clause for SBLOBs with
round robin inserting to all four sbspaces :-( )
15:58:04 Pagenum 2150646 (>= npused 2150583) is not free in bitmap
15:58:04 map 0x7fffff, bit 0x400000
15:58:04 Bitmap Pagenum 0x20be00, freemnum 4832
==========================

oncheck -ce output shows following errors:
==========================
Validating extents for Space 'sbspace1' ...

Chunk Pathname Pagesize(k) Size(p)
Used(p) Free(p)
ERROR: Missing pages 11930126 through 11930126 inclusive

Added missing smart object pages back to chunk
ERROR: Missing pages 14376887 through 14376887 inclusive
==========================

Other oncheck are without errors.

From user point of view some of transactions that inserts/updates data
with SBLOB are ok, some are rolled back - it can be viewed in logical logs
Could you please advice me how to fix this issue or help me with some
workarround?

Thanks, lempo

Messages In This Thread

• Bitmap is wrong for partp in sbspace
  Peter Lempochner -- Wednesday, 7 August 2013, at 9:51 a.m.
  • Re: Bitmap is wrong for partp in sbspace
    Art Kagel -- Wednesday, 7 August 2013, at 11:10 a.m.

IDS Forum is maintained by Administrator with WebBBS 5.12.