| 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 |
I have a customer running DBMS 6.1-0. Lately they have notice the system
getting slow... users are complaining. They checked out the user stall
messages screen and see that there are some
"waiting for TSN block 2 (CR)" stalls... usually, when things are running
fine, these are not present on the stall screen.
This is what the stall screen looks like:
Node: SPRT02 DEC DBMS V6.1-0 Performance Monitor 13-MAY-1997 08:58:08
Rate: 3.00 Seconds Active User Stall Messages Elapsed: 00:31:44.98
Page: 1 of 4 DSA3:[MMS_PROD.DBMS]MOS.ROO;2 Mode: Online
--------------------------------------------------------------------------------
Process.ID Since...... Stall.reason............................. Lock.ID.
2041AEB4:1 writing ROOT file
20414F42:1 waiting for TSN block 2 (CR) 6B003C39
20410F80:1 writing ROOT file
2040F06B:1 writing ROOT file
20411550:1 writing ROOT file
20412A5B:1 08:58:08.23 - reading pages 7:59614 to 7:59620
2040DFA4:1 writing ROOT file
2041434C:1 writing ROOT file
20417ABC:1 writing ROOT file
20410056:1 writing ROOT file
20417E81:1 writing ROOT file
2040E36D:1 writing ROOT file
204121B8:1 writing ROOT file
2041216F:1 writing ROOT file
20416A48:1 writing ROOT file
20410446:1 waiting for TSN block 2 (CR) 50001EC4
2041909E:1 writing ROOT file
--------------------------------------------------------------------------------
Display_menu Exit Help LockID >next_page <prev_page Set_rate Write_screen !
I had her check for DBR processes, there are none running when this happens.
Fast Commit, CTJ are disabled.
Running with 1 journal file.
Anything else I can get you???
What might be wrong and what can they do to improve the situation... it is
happening more and more frequently.
I had her send me the following stat screens too, not sure if they will
help you or not:
Node: SPRT02 DEC DBMS V6.1-0 Performance Monitor 13-MAY-1997 08:59:17
Rate: 3.00 Seconds Active User Stall Messages Elapsed: 00:32:54.00
Mode: Online
+--Lock Information: 6B003C39----------------------------------+ --------------
� <-- MORE � Lock.ID.
� Owner: 20414F42 Fin->Folio 6B003C39 CR NL Grant �
� Owner: 2041815D PM5 1B005723 CR NL Grant � 6B003C39
� Owner: 20410F32 PICK SHEET PRT4 160036C1 CR NL Grant �
� Owner: 20415EB0 BILLING CLERK 40005D64 EX NL Grant �
� Owner: 20205E54 John Thompson 3B0030C5 NL NL Grant �
� Owner: 2041276C _TNA2247: 2000440D EX NL Grant � ion
� Owner: 20412B74 E CZLAPINSKI 010063E4 NL NL Grant �
� Owner: 2020012B DBMS_MONITOR 0F004261 NL NL Grant �
� Owner: 20410522 Julie Spice 7A004084 NL NL Grant �
� Owner: 2040A4AE IP Roll RF 4 6500608E EX NL Grant �
� Owner: 2040F535 MHLD 310040E8 NL NL Grant �
� Owner: 2041434C IP Roll RF 2 0C006168 CR NL Grant �
� Owner: 2040DFA4 Fin->JAG2 360021A7 NL NL Grant �
� Owner: 20411C72 Mat Hndlr LdrHH 52000D2F EX NL Grant �
� MORE --> �
+--------------------------------------------------------------+ 50001EC4
--------------------------------------------------------------------------------
Use prev/next or up/down keys to change pages; any other key to exit display
Node: SPRT02 DEC DBMS V6.1-0 Performance Monitor 13-MAY-1997 09:00:30
Rate: 3.00 Seconds File IO Statistics Elapsed: 00:34:07.12
Page: 1 of 1 DSA3:[MMS_PROD.DBMS]MOS.ROO;2 Mode: Online
--------------------------------------------------------------------------------
For File: All data/snap files
statistic........... rate.per.second............. total....... average......
name................ max..... cur..... avg....... count....... per.trans....
total I/Os 0 1711776 22.3
(Synch. reads) 117 0 6.4 1125265 14.6
(Synch. writes) 3 0 0.1 34402 0.4
(Extends) 0 0 0.0 11 0.0
(Asynch. reads) 340 0 2.0 360131 4.6
(Asynch. writes) 15 0 1.0 191967 2.5
statistic........... blocks.transferred......... stall.time.(x100)...........
name................ avg.per.I/O.. total........ avg.per.I/O... total........
total I/Os 11.2 19324483 1.5 2657020
(Synch. reads) 12.4 13988216 1.9 2212351
(Synch. writes) 2.1 74786 2.0 70634
(Extends) 200.7 2208 15.0 165
(Asynch. reads) 13.4 4853684 0.3 109707
(Asynch. writes) 2.1 405589 1.3 264163
--------------------------------------------------------------------------------
Display_menu Exit Help Options Reset Set_rate Write_screen !
Node: SPRT02 DEC DBMS V6.1-0 Performance Monitor 13-MAY-1997 09:01:34
Rate: 3.00 Seconds Summary Locking Statistics Elapsed: 00:35:11.00
Page: 1 of 1 DSA3:[MMS_PROD.DBMS]MOS.ROO;2 Mode: Online
--------------------------------------------------------------------------------
statistic......... max. cur. 10 20 30 40 50
name.............. rate rate +-------------------------------------------------+
� � � � � �
locks requested 3165 0 � � � � � �
rqsts not queued 87 0 � � � � � �
rqsts stalled 2 0 � � � � � �
rqst timeouts 0 0 � � � � � �
rqst deadlocks 0 0 � � � � � �
locks promoted 450 0 � � � � � �
proms not queued 11 0 � � � � � �
proms stalled 9 0 � � � � � �
prom timeouts 0 0 � � � � � �
prom deadlocks 0 0 � � � � � �
locks demoted 80 0 � � � � � �
locks released 3176 0 � � � � � �
blocking ASTs 19 0 � � � � � �
stall time x100 19 0 � � � � � �
+-------------------------------------------------+
--------------------------------------------------------------------------------
Display_menu Exit Help Numbers Options Reset Set_rate Time_plot Write_screen Yan
Node: SPRT02 DEC DBMS V6.1-0 Performance Monitor 13-MAY-1997 09:02:21
Rate: 3.00 Seconds AIJ Statistics Elapsed: 00:35:57.86
Page: 1 of 1 DSA3:[MMS_PROD.DBMS]MOS.ROO;2 Mode: Online
--------------------------------------------------------------------------------
statistic......... max. cur. 10 20 30 40 50
name.............. rate rate +-------------------------------------------------+
� � � � � �
AIJ file writes 1 0 � � � � � �
data 1 0 � � � � � �
control 0 0 � � � � � �
file extend 0 0 � � � � � �
switch over 0 0 � � � � � �
records written 18 0 � � � � � �
blocks written 9 0 � � � � � �
filler bytes 40 0 � � � � � �
group commits 1 0 � � � � � �
cache overflows 0 0 � � � � � �
quick flushes 0 0 � � � � � �
ARB pool searches 6 0 � � � � � �
pool empty 0 0 � � � � � �
lock rebuilds 0 0 � � � � � �
AIJ file reads 0 0 � � � � � �
-------------------------------------------------------------------------------+
Display_menu Exit Help Numbers Options Reset Set_rate Time_plot Write_screen Yan
Node: SPRT02 DEC DBMS V6.1-0 Performance Monitor 13-MAY-1997 09:02:45
Rate: 3.00 Seconds AIJ Information Elapsed: 00:36:21.87
Page: 1 of 1 DSA3:[MMS_PROD.DBMS]MOS.ROO;2 Mode: Online
--------------------------------------------------------------------------------
Journaling: Enabled Shutdown: 60 Notify: Disabled State: Accessible
ALS: Manual ABS: Disabled ACE: Disabled FC: Disabled CTJ: Disabled
After-Image.Journal.Name....... SeqNum AIJsize CurrEOF Status. State.......
MOS 520 88064 87118 Current Accessible
--------------------------------------------------------------------------------
Display_menu Exit Help >next_page <prev_page Refresh Set_rate Write_screen !
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 2631.1 | my guess | HOTRDB::LASTOVICA | Can you be a closet claustrophobic? | Tue May 13 1997 16:31 | 6 |
your screen display doesn't show any active stalls except for
"20412A5B:1 08:58:08.23 - reading pages 7:59614 to 7:59620".
*sometimes* if there is massive root file I/O after a cluster node
fails (paul will correct me if I'm wrong; I think that's the trigger of
the case), you need to close and reopen the database. I believe that
this problem has been fixed (at least in Rdb).
| |||||
| 2631.2 | HOTRDB::PMEAD | Paul, [email protected], 719-577-8032 | Tue May 13 1997 18:15 | 6 | |
Oh yeah, thanks for the reminder Norm.
Look and see if there is a "lot" of root-file I/O. If so then it is
likely to be that problem. I don't remember exactly when it was fixed,
but I think it was V6.1-11. Look for a release note with a title
something like "Excessive root file I/O".
| |||||
| 2631.3 | Thanks | BROKE::BASTINE | Tue May 13 1997 20:02 | 7 | |
I remember seeing that release note on the Rdb side... I'll contact the customer again and see if a node crashed prior to the problem... I'll also have her upgrade... Thanks, Renee | |||||
| 2631.4 | HOTRDB::LASTOVICA | Can you be a closet claustrophobic? | Tue May 13 1997 20:08 | 3 | |
you might want to determine if the slowdown had anything to do with
massive I/O to the root file. That'll be your clue that this might be
the problem that paul and I are thinking of.
| |||||
| 2631.5 | HOTRDB::PMEAD | Paul, [email protected], 719-577-8032 | Wed May 14 1997 10:54 | 2 | |
It doesn't have to be a node crash -- a process failure can trigger the
problem also.
| |||||