[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
| Title: | VAX DBMS |
| Notice: | THIS NOTESFILE IS NOT A FORMAL SUPPORT CHANNEL |
| Moderator: | SCARY::CHARLAND |
|
| Created: | Thu Feb 20 1986 |
| Last Modified: | Tue Jun 03 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2642 |
| Total number of notes: | 11044 |
2609.0. "Page info belongs to a different page of an entirely different storage area" by svrav1.au.oracle.com::AMENDOZA () Fri Mar 14 1997 06:04
Hi,
I came across a very interesting scenario in one of our DBMS customer here
in Asia. They are running DBMS V4.2A on a 2 node VAX cluster with VMS
V5.5.2. It looks like a classic case of a caching problem to me but I
would like to get some idea on how the problem happened so that I can
explain it to the customer.
I will summarize here the facts that I gathered so far:
1. The problem was reported to me when they are consistently getting
bugcheck when certain records from their database is being accessed.
exception: PIO$FETCH_RET + 208
when dbkey 13:79449:61930 is processed by the PIO$FETCH_RET routine.
2. Solution applied at that time is to restore from the previous backup.
- we backed up the corrupt db first before restoring the previous
backup for further analysis.
3. Db has buffer count of 32, each is 16 blocks long. Area 13 (VENAREA)
is uniform page format with page size of 2 blocks. When I look in
the IOCB context of the dump for page 13:79449, I found it to be
thrashed. Content is:
0000 00000000 0000 page 0, physical area 0
00000000 0006 checksum =3D 00000000
00000000 00000000 000A time stamp =3D 17-NOV-1858 00:00:=
00.00
0000 0000 00120 free bytes, 0 locked
0000 0016 0 lines
00000000000000000000000000000000 0018 ** junk ** '................'
:::: (61 duplicate lines)
0000000000000000 03F8 ** junk ** '........'0C
Pages 79447, 79448, 79450, 79451, 79452, 79453 and 79454 read in the
same db buffer are fine.
4. Corrupt db backup is restored on another location.
DBO/dump/area=VENAREA/start=79499/end=79499 gives a different page
belonging to an entirely different area of the database. Again,
the adjacent pages (79447, 79448, 79450 etc.) are alright as before.
0036 00004F30 0000 page 20272, physical area 54
9DB34E1B 0006 checksum =3D 9DB34E1B
009ADE72 99D8C7C0 000A time stamp =3D 5-JAN-1997 09:23:=
40.22
.................
Note: I omitted the details of dump here for brevity but included them
in the attachments. Page 20272 belongs to APACTAREA.
In short: DBO/dump/area=VENAREA/start=79499/end=79499 and
DBO/dump/area=APACTAREA/start=20272/end=20272
gives the same 2 blocks worth of information. VENAREA page size
is 2 blocks while APACTAREA is 3 blocks.
We verified that the dump we are getting for page 13:79499 is
exactly the same thing by going to dbo/alter and displaying the
page itself.
5. The customer disk configuration:
- Disk Controller - HSD10
- Disk type where VENAREA and APACTAREA sitting - RZ29 (4.3 GB)
- (128 x 2)MB Memory - caching done in memory
- Clustered System with 2 nodes, diagram as below
NODE 1 NODE 2
(WDMA01) (WDMA02)
| |
| |
+-------------+
|
HSD10
|
|
------------------------
| | |
RZ29 RZ28 ..........6 DEVICES
Question:
---------
1. Is the classic case of caching problem?
2. What would be the probable mechanism on how this kind of problem
was arrived?
thanks,
albert.
--------------------------------------------------------------------
Attachments:
A. IOCB context info from bugcheck dump for 79499's clump:
000D 00013657 0000 page 79447, physical area 13
F80901BF 0006 checksum =3D F80901BF
009AF9B7 C05A0F20 000A time stamp =3D 9-FEB-1997 02:16:41.49
0000 00B0 0012 176 free bytes, 0 locked
0003 0016 3 lines
.....................
page detail omitted
.....................
000D 00013658 0000 page 79448, physical area 13
30BD3BE9 0006 checksum =3D 30BD3BE9
009AEA69 A5C72B00 000A time stamp =3D 20-JAN-1997 14:49:48.72
0000 0150 0012 336 free bytes, 0 locked
0002 0016 2 lines
.....................
page detail omitted
.....................
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
! page 79449 should be here but instead a "thrashed"/"severely !
! corrupted" page has been retrieved from the database. !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
0000 00000000 0000 page 0, physical area 0
00000000 0006 checksum =3D 00000000
00000000 00000000 000A time stamp =3D 17-NOV-1858 00:00:=
00.00
0000 0000 00120 free bytes, 0 locked
0000 0016 0 lines
00000000000000000000000000000000 0018 ** junk ** '................'
:::: (61 duplicate lines)
0000000000000000 03F8 ** junk ** '........'0C
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
! end of "trashed page" !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
000D 0001365A 0000 page 79450, physical area 13
B9F2D681 0006 checksum =3D B9F2D681
009AE3B6 2C84E160 000A time stamp =3D 12-JAN-1997 02:09:58.39
0000 0160 0012 352 free bytes, 0 locked
0002 0016 2 lines
.....................
page detail omitted
.....................
000D 0001365B 0000 page 79451, physical area 13
A319465E 0006 checksum =3D A319465E
009AF9B7 C05A0F20 000A time stamp =3D 9-FEB-1997 02:16:41.49
0000 00B2 0012 178 free bytes, 0 locked
0004 0016 4 lines
.....................
page detail omitted
.....................
000D 0001365C 0000 page 79452, physical area 13
40465947 0006 checksum =3D 40465947
009ADE1D FB34CB40 000A time stamp =3D 4-JAN-1997 23:17:56.34
0000 014C 0012 332 free bytes, 0 locked
0002 0016 2 lines
.....................
page detail omitted
.....................
000D 0001365D 0000 page 79453, physical area 13
E5A6B2A0 0006 checksum =3D E5A6B2A0
009AE75E 3ABDCB60 000A time stamp =3D 16-JAN-1997 17:50:31.19
0000 0154 0012 340 free bytes, 0 locked
0002 0016 2 lines
.....................
page detail omitted
.....................
000D 0001365E 0000 page 79454, physical area 13
70F94139 0006 checksum =3D 70F94139
009AFE72 D51FD100000A time stamp =3D 15-FEB-1997 02:45:56.88
0000 0120 0012 288 free bytes, 0 locked
0002 0016 2 lines
.....................
page detail omitted
.....................
B. From restored copy of corrupted database:
DBO/DUMP/AREA=VENAREA/START=79499/END=79499
*------------------------------------------------------------------------=
------
* VAX DBMS V4.2A-00 13-MAR-1997 16:16=
:42.59
*
* Dump of storage area VENAREA
* Filename: $DISK11:[MANMAN.MANDB268]VENAREA.DBS;1
* Database: $DISK11:[MANMAN.MANDB268]MANDB.ROO;1
*
*------------------------------------------------------------------------=
------
0036 00004F30 0000 page 20272, physical area 54
9DB34E1B 0006 checksum =3D 9DB34E1B
009ADE72 99D8C7C0 000A time stamp =3D 5-JAN-1997 09:23:=
40.22
0000 00C0 0012 192 free bytes, 0 locked
0007 0016 7 lines
0005 05F6 0018 line 0: offset 05F6, 5 bytes
00DE 0518 001C line 1: offset 0518, 222 bytes
00BB 045C 0020 line 2: offset 045C, 187 bytes
00D3 0388 0024 line 3: offset 0388, 211 bytes
00DB 02AC 0028 line 4: offset 02AC, 219 bytes
00DF 01CC 002C line 5: offset 01CC, 223 bytes
00D7 00F4 0030 line 6: offset 00F4, 215 bytes
00000000000000000000000000000000 0034 free space '................'
:::: (11 duplicate lines)
009B 00F4 line 6: record type 155
07 0072 00F6 114 bytes in 7 sets/dynamic items
0002 80 0017 00F9 23 bytes, data item type 2
05 00FE unlike-run of 6 bytes of data
323230313831 00FF data '181022'
30 83 0105 like-run of 4 bytes of '0'
05 0107 unlike-run of 6 bytes of data
31303038352D 0108 data '-58001'
20 87 010E like-run of 8 bytes of ' '
0008 80 0007 0110 7 bytes, data item type 8
20 97 0115 like-run of 24 bytes of ' '
000A 80 0023 0117 35 bytes, data item type 10
1A 011C unlike-run of 27 bytes of data
44414F4C444145482045525554584946 011D data 'FIXTURE HEADLOAD'
56204E4F4547525554532C 012D data ',STURGEON V'
20 82 0138 like-run of 3 bytes of ' '
0010 80 000C 013A 12 bytes, data item type 16
03 013F unlike-run of 4 bytes of data
9000482B 0140 data '+H..'
00 83 0144 like-run of 4 bytes of '.'
0012 80 000A 0146 10 bytes, data item type 18
01 014B unlike-run of 2 bytes of data
4120 014C data ' A'
00 85 014E like-run of 6 bytes of '.'
0151 0F 0150 15 bytes, storage set type 337
00BA37 19 0153 next 54:67943:6
13087B 1A 0157 owner 54:98231:1
0DB9FC 15 015B prior 54:10959:2
0166 0B 015F 11 bytes, storage set type 358
0F 01 0162 next 54:20272:5
4E840C 42 0164 owner 58:14192:2
CC 06 0168 prior 54:20268:2
.... 97 bytes of static data
3830202020203039363534310000FF03 016A data '....145690 08'
00004000478901309189013091893639 017A data '96..0...0..G.@..'
20202020202000000000000100010000 018A data '.......... '
00000000000000000020202020202020 019A data ' .........'
00000000000000000000000000000000 01AA data '................'
:::: (1 duplicate line)
00 01CA data '.'
00 01CB padding '.'
009B 01CC line 5: record type 155
08 007A 01CE 122 bytes in 8 sets/dynamic items
0002 80 0018 01D1 24 bytes, data item type 2
02 01D6 unlike-run of 3 bytes of data
313132 01D7 data '211'
30 82 01DA like-run of 3 bytes of '0'
09 01DC unlike-run of 10 bytes of data
31303038352D30303231 01DD data '1200-58001'
20 87 01E7 like-run of 8 bytes of ' '
0008 80 0007 01E9 7 bytes, data item type 8
20 97 01EE like-run of 24 bytes of ' '
000A 80 0019 01F0 25 bytes, data item type 10
10 01F5 unlike-run of 17 bytes of data
4E554F43434120504120524F444E4556 01F6 data 'VENDOR AP ACCOUN'
54 0206 data 'T'
20 8C 0207 like-run of 13 bytes of ' '
0010 80 000C 0209 12 bytes, data item type 16
03 020E unlike-run of 4 bytes of data
9000482B 020F data '+H..'
00 83 0213 like-run of 4 bytes of '.'
0012 80 000A 0215 10 bytes, data item type 18
01 021A unlike-run of 2 bytes of data
4120 021B data ' A'
00 85 021D like-run of 6 bytes of '.'
0151 0F 021F 15 bytes, storage set type 337
1266EE 1A 0222 next 54:95646:3
135CBC 1A 0226 owner 54:99579:1
119FEC 1A 022A prior 54:92462:1
0166 0D 022E 13 bytes, storage set type 358
4E840D 42 0231 next 58:14192:2
4E840D 42 0235 owner 58:14192:2
01 01 0239 prior 54:20272:6
0167 0F 023B 15 bytes, storage set type 359
4E840D 42 023E next 58:14192:2
4E840D 42 0242 owner 58:14192:2
4E840D 42 0246 prior 58:14192:2
.... 97 bytes of static data
3830202020203039363534310000FF03 024A data '....145690 08'
00004000478901309189013091893639 025A data '96..0...0..G.@..'
20202020202000000000FFFF00010000 026A data '.......... '
00000000000000000020202020202020 027A data ' .........'
00000000000000000000000000000000 028A data '................'
:::: (1 duplicate line)
00 02AA data '.'
00 02AB padding '.'
009B 02AC line 4: record type 155
0A 0076 02AE 118 bytes in 10 sets/dynamic item=
s
0002 80 0017 02B1 23 bytes, data item type 2
31 82 02B6 like-run of 3 bytes of '1'
0C 02B8 unlike-run of 13 bytes of data
31303038352D30303135343030 02B9 data '0045100-58001'
20 87 02C6 like-run of 8 bytes of ' '
0004 80 000A 02C8 10 bytes, data item type 4
03 02CD unlike-run of 4 bytes of data
003F7DE8 02CE data '=E8}?.'
0008 80 0007 02D2 7 bytes, data item type 8
20 97 02D7 like-run of 24 bytes of ' '
000A 80 0007 02D9 7 bytes, data item type 10
20 9D 02DE like-run of 30 bytes of ' '
0010 80 000C 02E0 12 bytes, data item type 16
03 02E5 unlike-run of 4 bytes of data
4000461C 02E6 data '.F.@'
00 83 02EA like-run of 4 bytes of '.'
0011 80 0008 02EC 8 bytes, data item type 17
01 02F1 unlike-run of 2 bytes of data
0002 02F2 data '..'
0012 80 000A 02F4 10 bytes, data item type 18
01 02F9 unlike-run of 2 bytes of data
4120 02FA data ' A'
00 85 02FC like-run of 6 bytes of '.'
0151 0E 02FE 14 bytes, storage set type 337
012CFF 19 0301 next 54:97327:4
03A79D 15 0305 owner 54:35241:1
581D 14 0309 prior 54:42829:4
0166 0D 030C 13 bytes, storage set type 358
0F 01 030F next 54:20272:3
4E840D 42 0311 owner 58:14192:1
4E840D 42 0315 prior 58:14192:1
015F 0D 0319 13 bytes, storage set type 351
0F 01 031C next 54:20272:3
1CB171 42 031E owner 55:6727:5
1CB171 42 0322 prior 55:6727:5
.... 97 bytes of static data
3430202020203437393533310000C810 0326 data '.=C8..135974 04'
00000000457A01306DA801306DA83639 0336 data '96=A8m0.=A8m0.zE....'
20202020202000000000FFFD00010000 0346 data '....=FD..... '
00000000000000000020202020202020 0356 data ' .........'
00000000000000000000000000000000 0366 data '................'
:::: (1 duplicate line)
00 0386 data '.'
00 0387 padding '.'
009B 0388 line 3: record type 155
09 006E 038A 110 bytes in 9 sets/dynamic items
0002 80 0018 038D 24 bytes,
C. Db and Storage area descriptions:
Database parameters
Root filename is $DISK9:[MANMAN.MANDB000]MANDB.ROO;1
Created at 8-JUL-1994 13:44:05.96
VAX DBMS structure level is 42.3
Maximum user count is 51
Maximum node count is 16
Buffer count is 32, each is 16 blocks long
DBR buffer count is 50, each is 16 blocks long
Adjustable record locking granularity factor count is 0
After image journaling is disabled
- AIJ default allocation is 0 blocks
- AIJ file extension is 512 blocks
No default recovery unit journal filename
Database is opened automatically on first user access
Snapshots are non-deferred (normal)
Release retrieval locks when no longer needed
Wait on record lock conflicts
Latest full backup file is dated 18-FEB-1997 00:15:14.40
Latest full backup transaction sequence number is 1794960
Database has never been incrementally restored
Latest full restore occured at 18-FEB-1997 02:52:49.29
Latest full verify occurred at 5-JAN-1997 14:06:58.18
Database has never been altered
Database has never been restructured
Storage area VENAREA
Area ID number is 13
Filename is $DISK9:[MANMAN.MANDB000]VENAREA.DBS;1
Page size is 2 blocks
Initial data page count was 100000
Current physical page count is 100391
Extends are enabled
- Area has never been extended
- Extend area by 20%, minimum of 20 pages, maximum of 4000 pages
- Volume set spreading is enabled
Area has space management pages
- Current SPAM page count is 391
- Interval is 256 data pages
- Thresholds are 70%, 85%, and 95%
Snapshots are not allowed
Area last backed up at 18-FEB-1997 00:15:14.40
Area has never been incrementally restored
Storage area APACTAREA
Area ID number is 54
Filename is $DISK9:[MANMAN.MANDB000]APACTAREA.DBS;1
Page size is 3 blocks
Initial data page count was 100000
Current physical page count is 100391
Extends are enabled
- Area has never been extended
- Extend area by 20%, minimum of 20 pages, maximum of 4000 pages
- Volume set spreading is enabled
Area has space management pages
- Current SPAM page count is 391
- Interval is 256 data pages
- Thresholds are 70%, 85%, and 95%
Snapshots are not allowed
Area last backed up at 18-FEB-1997 00:15:14.40
Area has never been incrementally restored
| T.R | Title | User | Personal
Name | Date | Lines
|
|---|