What is "reliable message" wait event in 10g??
Hi, all.
The database is 2 node RAC 10.2.0.2.0 on 32-bit windows 2003 EE SP1.
From time to time, I see "reliable message" wait event on the
"Top 5 Timed Events" section from AWR report.
The wait class is "Other".
There seems to be no manuals and html links explaining "reliable message"
wait event.
Thanks and Regards.
Message was edited by:
user507290
Message was edited by:
user507290
Metalink Doc 461052.1 is a fairly lightweight walk-through of a system resource profiling tool called LTOM, the "Lite Onboard Monitor". It has an appendix in which it simply points out "The following events were considered idle and ignored... reliable message..."
It also says that if you want to include something for wait analysis, just take it out of the list of idle and ignored events... so it's not exactly saying that the thing is an idle event, just that it was treated as such for the purposes of writing that document but that you might think differently.
On the other hand, this page:
http://download.oracle.com/docs/cd/B19306_01/server.102/b14237/waitevents003.htm
...is 10g Release 2's description of "wait events of greatest interest"... and "reliable message" isn't on it! From which one concludes, I suppose, that it isn't of great interest!
Similar Messages
-
What is "reliable message" wait event?
Hi, all.
What is "reliable message" wait event?
The wait class is "Other", not "Idle".
From time to time, I can see "reliable message" wait event
in the Top 5 Timed Events section from AWR reports.
In addition, what is "CPU time" in the Top 5 Timed Events section?
Thanks and Regards.
Message was edited by:
user507290Dear Yas.
Thanks for your reply.
The database is 2 node RAC database (10.2.0.2.0) on 32 bit windows 2003 EE SP1.
Does "AQ" mean Streams AQ???
Can I disable Streams AQ facility??
Thanks and Regards.
Message was edited by:
user507290
Message was edited by:
user507290 -
When there is archiving done to the recovery area, there is "rdbms ipc message" wait event spike.
What to conclude from it?
Wait Class Wait Event P1 Text P1 P2 Text P2 P3 Text P3 Wait Time (ms)
System I/O log file parallel write files 2 blocks 4096 requests 2 10
Idle rdbms ipc message timeout 150 0 0 90
System I/O log file parallel write files 2 blocks 4096 requests 2 10
Idle rdbms ipc message timeout 161 0 0 100
System I/O log file parallel write files 2 blocks 4096 requests 2 10
Idle rdbms ipc message timeout 181 0 0 190
System I/O log file parallel write files 2 blocks 4096 requests 4 10
Idle rdbms ipc message timeout 191 0 0 90
System I/O log file parallel write files 2 blocks 4096 requests 2 10
Idle rdbms ipc message timeout 201 0 0 90
ThanksYou can use the following details
Wait Class Wait Event P1 Text P1 P2 Wait Time(ms)
System I/O log file parallel write files 2 4096 10
Idle rdbms ipc message timeout 150 0 90
System I/O log file parallel write files 2 4096 10
Idle rdbms ipc message timeout 161 0 100
System I/O log file parallel write files 2 4096 10
Idle rdbms ipc message timeout 181 0 190
System I/O log file parallel write files 2 4096 10
Idle rdbms ipc message timeout 191 0 90
System I/O log file parallel write files 2 4096 10
Idle rdbms ipc message timeout 201 0 90 -
About single-task message wait event
Hello
I have several active users, some of them from 2 an 3 days ago with single-task message wait event and their last_call was many time ago. One of users, blocked to other users just a little while ago and Concurrency grow up to 20 % and was on increase. i had to kill this user and all were well, concurrency dessapeared.
How could I avoid this behavior??
How could I kill these type of user by automatic way??
ThanksHi,
according to me active user is different
according to me active user is session or oracle process is doing something else like dml or ddl execution or select stmt else session is idel.
trace the session what is doing? is may be problem with dead connection
use oradebug and set the 10046 at level 12 and format it with tkprof utility.
paste it thread
Kind Regards,
Rakesh jayappa -
What is "KJC: Wait for msg .." wait event in 10g??
Hi, all.
The database is 2 node RAC database (10.2.0.2.0)
on 32-bit windows 2003 EE SP1.
I found "KJC: Wait for msg sends to complete" wait event in
"Top 5 Timed Event" Section from AWR report.
What is "KJC: Wait for msg sends to complete" wait event??
The following is from UDUMP.
Dump file d:\oracle\product\10.2.0\admin\rac\udump\rac2_ora_5656.trc
Mon Sep 24 00:04:40 2007
ORACLE V10.2.0.2.0 - Production vsnsta=0
vsnsql=14 vsnxtr=3
Oracle Database 10g Enterprise Edition Release 10.2.0.2.0 - Production
With the Real Application Clusters, OLAP and Data Mining options
Windows Server 2003 Version V5.2 Service Pack 1
CPU : 4 - type 586, 2 Physical Cores
Process Affinity : 0x00000000
Memory (Avail/Total): Ph:5278M/8190M, Ph+PgF:6596M/10041M, VA:316M/2047M
Instance name: rac2
Redo thread mounted by this instance: 2
Oracle process number: 64
Windows thread id: 5656, image: ORACLE.EXE (SHAD)
*** 2007-09-24 00:04:40.156
*** ACTION NAME:() 2007-09-24 00:04:40.156
*** MODULE NAME:(OEM.SystemPool) 2007-09-24 00:04:40.156
*** SERVICE NAME:(RAC.world) 2007-09-24 00:04:40.156
*** CLIENT ID:() 2007-09-24 00:04:40.156
*** SESSION ID:(486.53) 2007-09-24 00:04:40.156
IPCSendMsg: could not initiate send on conn 0x5b0d3e98 to node [rac1 : 696 : 3996 : 359937], err 10054
IPCGetRequestInfo: failed a request rqh(0x5b060db8), type(6), status(2), bytes(0)
Thanks and Regards.
Message was edited by:
user507290
Message was edited by:
user507290This might have something to do with bug 5075434 - Small performance overhead in RAC (waits for "KJC: Wait for msg sends to complete").
Check metalink for further details. -
Wait event in 10g "events in waitclass others" , DB is stuck
Hi ,
I need help ,
no one can login into the DB , there are 5 updates running in parallel right now ,
only those who were inside before the updates began to run can work .
connect through listener is impossible , but without listener (i.e sqlplus user/password) is o.k .
in v$session_event for each update I see "events in waitclass others" wait event , with large values either in "total_waits" column , and "total_timeouts" .
I cannot find anything in asktom , or google .
yesterday we migrated from 9i to 10g .
a month ago we did the same in the dev area with no problems .
did anyone see that events before , and know what should be done to solve this problem ?
Thanks ,
Lior .sir,
my few parameters that i've set.
fast_start_mttr_target = 220
db_recovery_file_dest_size = 10G (includes backupset, archivelogs, flashback)
sga_target = 1048M
pga_aggregate_target = 600M
these two statements consuming significant time of wait..
SELECT TO_CHAR(OLDEST_FLASHBACK_TIME, 'YYYY-MM-DD HH24:MI:SS') FROM V$FLASHBACK_DATABASE_LOG
SELECT (:"SYS_B_0" - sum(percent_space_used)) + sum(percent_space_reclaimable)
FROM v$flash_recovery_area_usage
there is also lots of time consuming in control file sequential read..
any hints for the problem..which parameters i've to check or any other... -
Single-Task Message Wait Event
Discovered while researching some performance issues....
A statspack report on the metadata repository is showing a lot of single-task message waits.
Is this normal? What is the source of these waits?
Per Oracle 8i documentation... "Oracle is diminishing the support of single-task mode."
Why is this showing up in a 9.2.0.8 database???
Any insights welcome...
Thanks,
Henry
Edited by: OracleDBAinPA on Dec 8, 2009 6:56 AMAs you can see, no one replied to the message...
But, our performance issue was due to a memory leak in the listener (documented bug in 9.2.0.8).
Restarting the listener, also reduced the wait... -
Hello SAP Community,
I start by mentioning a few details about the system I'll be talking about in this subject:
- SAP NetWeaver 7.0
- Oracle Database 10.2g
I was reading the following Note: "Note 618868 - FAQ: Oracle performance", in order to try to understand what's causing the oracle database to have slow performance.
While reading section 3 "How can I determine whether the general database performance can be optimized?" I found out that the ratio of "Busy wait time to CPU time" is away above the recommended 60:40 value. I'm getting a 94:6 ratio. This value was calculated using the query:
SELECT
ROUND((STM1.VALUE - STM2.VALUE) / 1000000) "BUSY WAIT TIME (S)",
ROUND(STM2.VALUE / 1000000) "CPU TIME (S)",
ROUND((STM1.VALUE - STM2.VALUE) / STM1.VALUE * 100) || ' : ' ||
ROUND(STM2.VALUE / STM1.VALUE * 100) RATIO
FROM V$SYS_TIME_MODEL STM1, V$SYS_TIME_MODEL STM2
WHERE STM1.STAT_NAME = 'DB time' AND STM2.STAT_NAME = 'DB CPU';
With such high values, SAP recommends to improve system performance doing some "wait event tuning".
Can someone give me some directions about this subject? Some guides specific to this subject would be nice. Any further information about my system you may require, please ask me.
Thanks in advance.
Best regards,
Daniel GarridoHello again,
Before I did any changes to the Oracle's parameters I checked the Note 619188 - FAQ: Oracle wait events, to understand what could be causing such high event wait time.
With the query:
SELECT EVENT, TOTAL_WAITS, TIME_WAITED, AVG_MS,
ROUND(RATIO_TO_REPORT(TIME_WAITED) OVER () * 100) PERCENT
FROM (SELECT SUBSTR(EVENT, 1, 30) EVENT, TOTAL_WAITS, TIME_WAITED,
ROUND(TIME_WAITED_MICRO / TOTAL_WAITS / 1000, 2) AVG_MS
FROM V$SYSTEM_EVENT
WHERE WAIT_CLASS NOT IN ('Idle', 'System I/O')
UNION
SELECT 'CPU' EVENT, NULL, VALUE, NULL
FROM V$SYSSTAT
WHERE STATISTIC# = 12
ORDER BY 3 DESC)
WHERE ROWNUM <=10;
I got the non-idle events that took more time in my system and the result was:
Result of the SELECT statement
EVENT
TOTAL_WAITS
TIME_WAITED
AVG_MS
PERCENT
log file switch (archiving nee
578.686
57.850.863
999.69
80
buffer busy waits
712.163
6.420.932
90.16
9
CPU
0
2.791.238
4
db file sequential read
4.005.546
1.746.442
4.36
2
log file sync
10.176.490
1.577.177
1.55
2
enq: TX - row lock contention
854.451
642.955
7.52
1
db file scattered read
1.055.533
621.332
5.89
1
enq: CF - contention
210.085
246.910
11.75
0
read by other session
561.558
119.910
2.14
0
log file switch completion
10.777
85.843
79.65
0
So most of the TIME_WAITED for wait events was because of the "log file switch (archiving needed)", after reading what could cause such wait event, I understood this was related with a problem I previously had in the server, where the archiving folder was with no space left. (Meanwhile the backup of the archives is being done and so the folder is being cleaned on a daily basis).
Thank you all for your help! -
Hi all,
We have a 2 node RAC 10.2.0.3 database.From quite sometime we are experiencing "transaction" wait event apart from "enq:TX".
What is this "transaction wait event"? How does it differ from enq:TX row lock contention event. I have searched a lot but couldn't find any proper document to describe it.
Has any one encountered this event before?
Thanks.Yasir Hashmi wrote:
Hi all,
We have a 2 node RAC 10.2.0.3 database.From quite sometime we are experiencing "transaction" wait event apart from "enq:TX".
What is this "transaction wait event"? How does it differ from enq:TX row lock contention event. I have searched a lot but couldn't find any proper document to describe it.
Has any one encountered this event before?
Thanks.what SQL produces this event?
against which object type does this wait occur?
do as below so we can know complete Oracle version & OS name.
Post via COPY & PASTE complete results of
SELECT * from v$version; -
Streams AQ wait event on Oracle 10g
Hello,
I have ECC 6.0 on W2k3 with Oracle. I have some wait event about Streams AQ :
Streams AQ: waiting for messages in the queue
Streams AQ: qmn slave idle wait
Streams AQ: qmn coordinator waiting for slave to start
What does it mean ? What can I do to fix that?
From what I read, it's seems to have something to do with parameter : aq_tm_processes
What this parameter whould be set to? It seems to be set to O now.
Thank you for any help,
NicholasHi,
What is the Patch Level of Oracle 10g which is in use ?
Please refer Oracle Meta link 428441.1 to get more information. It will tell you the reason and the possible alternatives to deal with it. You can refer SAP Note 758563 to get Oracle Meta link access.
Unless you use Oracle Streams Advanced Queuing , there's no need to set this parameter.
If AQ_TM_PROCESSES is not specified or is set to 0, then the queue monitor is not created.
In 10gR2 parameter AQ_TM_PROCESSES shouldn't be set explicitly in pfile/spfile, because Oracle autotunes it.
Also refer the [this link|SRM Alert Management does not determine recipient runtime?; to get more info.
Regards,
Bhavik G. Shroff -
Reliable message - Top first timed events
Hi All,
One fo my Db having reliable message as top first timed events. Does anyone have an idea about this cause and the solution for this . Please do the needful.
Event Waits Time(s) Avg wait (ms) % DB time Wait Class
reliable message 685,139 24,625 36 29.58 Other
Thankshttp://arulselvaraj.blogspot.com/2011/01/drop-tablespace-waiting-on-reliable.html
-
What is ges reusing os pid wait event
What is wait event "ges reusing os pid". In our RAC environment it is one of the top wait events. How to minimze it.
This is a wait event in Oracle 10g for Global Enqueue Services (ges) waiting on an operating system process id (os pid).
How to resolve this issue? I checked the bug list on Metalink and there is a patch set for the issue that may help.
Question: what version and patch release are you running for Oracle RAC?
Also, you probably want to tune your public network and private interconnects between the nodes in your Oracle RAC cluster.
Regards,
Ben Prusinski
http://oracle-magician.blogspot.com/ -
What is "ges enter server mode" wait event..
I cannot find 'ges enter server mode' wait event definition.
What is it?10g onwards, Instance recovery is done in two phases. First phase scans the blocks to be recoverd and applied from rdo log files and the second phase actually does that.
In a RAC instance the during the instance recovery, first pass scan can be delayed by 300ms-1.5s waiting on GRD (Global resource directory). During that time SMAON enters the wait event which is called "ges enter server mode" . -
What's wrong from this wait event
Please,
Below are two tables showing respectively database wait event by wait class and session wait event by wait class.
1. Database wait event by wait class
WAIT TOTAL PCT TIME PCT
CLASS WAITS WAITS WAITED_SECS TIME
Application 7427 .08 1572.45 76.29
User I/O 50416 .57 193.24 9.38
Network 8714874 98.66 177.67 8.62
System I/O 48169 .55 85.98 4.17
2. Session wait event by wait class
SID USER WAIT TOTAL TIME_
NAME CLASS WAITS WAITED_SECS
318 PMS1000 Application 110 321.64
259 PMS1000 Application 81 212.8
318 PMS1000 Network 541943 31.8
259 PMS1000 Network 258368 17.76
258 PMS1000 Network 132774 9.34
318 PMS1000 User I/O 1392 7.49
Top Events found:
CPU + WAIT for CPU
ROW lock contention
SQL*Net more data from/to client
Question:
What may cause the application wait_class to be at the top?, event though the row lock contention was found ?
I also think the system may sufering from a network bottleneck, I also thinking to set SDU parameter, but the network is 1Gb speed, and I don't know if this can help.
Does someone give me some clue to pinpoint what is going wrong wiht the above stats?
thanks enoughuser552326
I've used the "code" tags to make your first section of data more readable:
WAIT TOTAL PCT TIME PCT
CLASS WAITS WAITS WAITED_SECS TIME
Application 7427 .08 1572.45 76.29
User I/O 50416 .57 193.24 9.38
Network 8714874 98.66 177.67 8.62
System I/O 48169 .55 85.98 4.17 If you want to know what events belong to each wait class you can query v$event_name:
select wait_class, name
from v$event_name
order by wait_class, nameThe events in class "Application" are:
SQL*Net break/reset to client
SQL*Net break/reset to dblink
Streams capture: filter callback waiting for ruleset
Streams: apply reader waiting for DDL to apply
Wait for Table Lock
enq: KO - fast object checkpoint
enq: PW - flush prewarm buffers
enq: RO - contention
enq: RO - fast object reuse
enq: TM - contention
enq: TX - row lock contention
enq: UL - contention
As you can see, this is consistent with your comment about the top event being "row lock contention". The implication of the name given to this wait class is that it is your application design that is causing the problem. Your biggest problem is that your code allows your users to lock each other out.
Looking at the summary numbers, the time spent on waiting for other users to get out of the way is a very large fraction of your total wait - until you deal with that, problems relating to I/O and network appear to be pretty irrelevant. Having said that, you seem to do a very large number of round-trips on the network - so maybe the amount of time you are losing is not hugely significant compared to the amount of work you are getting done. (You didn't actually tell us how long it took or how many concurrent users, to accumulate this wait time).
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk -
"lms flush message acks" wait event
Dear All,
We are load testing our application in 10.2.0.1 database with 2 node RAC, in the AWR report top 5 wait event, we are getting "lms flush message acks" 90%.
I did search in google and metalink, but i could not able to find any related notes.
Please help..
Thanks,
Anand.[email protected] wrote:
We are load testing our application in 10.2.0.1 database with 2 node RAC, in the AWR report top 5 wait event, we are getting "lms flush message acks" 90%.
Something is waiting for the "other" node to acknowledge a "flush" message - so you need to look at the other node to see if you can see anything that might cause the flush message to see a slow response.
Of course, there may be congestion on the interconnect - but then various other RAC communications would also be slow - so it's more likely that the "flushing" is slow.
Reasons for flushing - we are telling the other node to clear part of its buffer cache, this might be related to frequent truncate commands (as the top of a shortlist). If you truncate an object, any dirty blocks for that object have to be written to disc, and any clean blocks have to be flushed from the cache; in a RAC environment the other nodes have to be told to do the same and your session has to wait for them to complete the write and flush.
In your case, you might check the code for frequency truncates - and check to see if you can see evidence for frequent slow write from dbwr (and also from lgwr) on the remote node.
Since you're running an early version of 10.2, I think problems of this type can even be related to truncates on global temporary tables due to some bugs that weren't fixed until 10.2.0.3. (And I think there were some problems with dynamic remastering in that version too, which caused similar flushing issues).
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk
To post code, statspack/AWR report, execution plans or trace files, start and end the section with the tag {noformat}{noformat} (lowercase, curly brackets, no spaces) so that the text appears in fixed format.
"Science is more than a body of knowledge; it is a way of thinking"
Carl Sagan
Maybe you are looking for
-
How to Install Win 7 ODBC for All Users?
I installed the Windows 7 64-bit ODBC drivers, but the install does not seem to have made the drivers available to all users, but rather only the installing user. This install is for an image that will be cloned as a virtual machine, so the drivers,
-
Okay, so basically I was really happy with the track, it has chopped up vocals all from the same vocal track. I was just experimenting and reversed a section of audio, but didn't save the changes because i didn't want to save this reverse. however no
-
Mandatory creation profile on scheduling agreements
Hello Gurus, I need to put a value on creation profile field on schedulling agreements. I have to choices: - Put the value on supplier data. It is a plant-dependant value i my client doesn't want to create this data. - Made it mandatory. Th
-
Panasonic TC-P50S2 In need of HELP
HI ALL- I received Sony MDRRF970RK wireless headphones. It seems my TV has no headphone jack. I've been using a Samsung Soundbar HW-C450 I tried plugging the headphones into that jack & they didn't work. I talked to someone at Panasonic & he said "
-
Long-running Batch eWay JCD - threading and stopping
I have a requirement to poll for files on a remote server, every few seconds, almost continuously. Furthermore, without making a connections every polling interval. In other words, connect once and stay running, If a file is found, send a JMS message