RMAN slow on exit
I have my backup destination as an NFS mount. My environment currnetly is linux to linux, red hat. I hvae 10.2.0.3.
The performance seems real bad, backup seems to take a really long time. The worst is, even when I just do
RMAN> list recoverable backup;
It takes a really long time.
The worst is, even if I exit rman, it takes a really long time to get to the command prompt.
Is there something in the way I'm doing a nfs mount?
mount -t nfs -o rw,bg,intr,hard,timeo=600,wsize=32768,rsize=32768,nfsvers=3,tcp back.testy.com:/data/back_test /home/oracle/backup_mount
Edited the subject since it turned out to be a bug in RMAN and this problem had nothing to do with NFS
I guess there are a couple of bugs on this issue.
Bug 5047758 - RMAN SLOW PERFORMANCE ON EXIT (UPDATING STATUS STATISTICS)
Version affected : 10.1
Version fixed : 10.2 due to base Bug 3325292 OPTIMIZER NOT ABLE TO USE INDEXES IN QUERY USING V$SQL_PLAN
Diagnosis : Expensive statement =
select round(sum(MBYTES_PROCESSED)) ,round(sum(INPUT_BYTES)),
round(sum(OUTPUT_BYTES))
from V$RMAN_STATUS
start with (RECID=:b1 and STAMP=:b2)
connect by prior RECID=parent_recid
Workaround : RMAN>sql "alter session set optimizer_mode=RULE";
Remark : It might be that Bug 5247609 is a followup / new occurence of this bug
Bug 5247609 RMAN SLOW PERFORMANCE DURING REGISTER DATABASE/OPEN RESETLOGS
Version affected : 10.2
Version fixed : 10.4
Diagnosis : Expensive statement =
select round(sum(MBYTES_PROCESSED)) ,round(sum(INPUT_BYTES)),
round(sum(OUTPUT_BYTES))
from V$RMAN_STATUS
start with (RECID=:b1 and STAMP=:b2)
connect by prior RECID=parent_recid
Workaround : RMAN>sql "alter session set optimizer_mode=RULE";So I did the same
run
sql "alter session set optimizer_mode=RULE";
BACKUP ARCHIVELOG ALL TAG = '${TAG_TO_USE}' FORMAT = '${FILE_LOCATION}/${FILE_NAME}' DELETE INPUT;
}Got the prompt back in a few minutes as soon as the backup was done.
Similar Messages
-
RMAN script doesnt exit shell after backup
Hi,
I've the following script on AIX 6.1, Oracle 11.2.0.1.0:
export ORACLE_HOME=/oracle/medprod/product/11.2.0/dbhome_1
export ORACLE_SID=MEDPROD
export PATH=$ORACLE_HOME/bin:$PATH
export LOG_DIR=/home/medinous/backup/logs
export LOG_FILE=$LOG_DIR/medprod_daily_backup_`date +%d%m%Y%X`.log
rman target / catalog rman/rman@rman log=$LOG_FILE <<EOF
execute script medinous_daily;
exit;
EXIT;
EOF
and:
printing stored script: medinous_daily
+{+
allocate channel t1 type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=/usr/tivoli/tsm/client/oracle/bin64/tdpo.opt)';
allocate channel t2 type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=/usr/tivoli/tsm/client/oracle/bin64/tdpo.opt)';
allocate channel t3 type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=/usr/tivoli/tsm/client/oracle/bin64/tdpo.opt)';
backup format 'medinous_dailybackup_%d_%s_%t' database;
backup format 'medinous_dailyarchbackup_%d_%s_%t' archivelog all not backed up 1 times;
crosscheck backup;
delete noprompt obsolete;
delete noprompt expired backup;
delete noprompt expired archivelog all;
release channel t1;
release channel t2;
release channel t3;
+}+
For some reason, the rman script, scheduled via crontab doesn't exit after successful backup, the process just remain alive, even though there's an explicit exit command at the end.
Any idea?
dulaNo, the script actully executes successfully. The problem is that after executing, the process does'nt die, it doesnt exit rman:
channel t1: finished piece 1 at 28-DEC-10
piece handle=prodarchbackup_PROD_666_738921153 tag=TAG20101228T075233 comment=API Version 2.0,MMS Version 5.5.2.0
channel t1: backup set complete, elapsed time: 00:00:03
channel t2: finished piece 1 at 28-DEC-10
piece handle=prodarchbackup_PROD_667_738921153 tag=TAG20101228T075233 comment=API Version 2.0,MMS Version 5.5.2.0
channel t2: backup set complete, elapsed time: 00:00:03
channel t3: finished piece 1 at 28-DEC-10
piece handle=prodarchbackup_PROD_668_738921153 tag=TAG20101228T075233 comment=API Version 2.0,MMS Version 5.5.2.0
channel t3: backup set complete, elapsed time: 00:00:03
Finished backup at 28-DEC-10
Starting Control File and SPFILE Autobackup at 28-DEC-10
piece handle=c-148294895-20101228-00 comment=API Version 2.0,MMS Version 5.5.2.0
Finished Control File and SPFILE Autobackup at 28-DEC-10
released channel: t1
released channel: t2
released channel: t3
*<< I expect the rman shell script to exit at this point and 'die' >>>>* -
Rman hangs at exit if locally connected
hiho,
i've a nasty problem. if i connect rman locally 'rman nocatalog target /' it will
hang at exit.
i tried to connect at rman-prompt with the same result.
im using 8.1.6.1.0 on linux.
# su - oracle
[oracle]$ rman nocatalog
Recovery Manager: Release 8.1.6.1.0 - Production
RMAN> connect target;
RMAN-06005: connected to target database: ACCSRV02 (DBID=1987850870)
RMAN-06009: using target database controlfile instead of recovery catalog
RMAN> exit
Recovery Manager complete.
rman hangs
can anyone confirm this? workaround? any
hints?
thanks for your time,
ciao -ap
nullHi!
I got same results on linux RH6.2 with 8.1.6.1.0, with 8.1.7 work fine ;)
You're right!
A10! -
My keynote is very slow when exiting from a slideshow, why
Anyone else suffering from keynote acting very slowly when exiting a slideshow?
Yes it's slow in Mavericks, not in other OS versions.
The software is only weeks old so there are many issues with both products. -
I'm running Eclipse on a Win XP desktop w/ 384mb memory. The Win and Java cache are probably 1 gig each. The app runs slow as you can imagine. Even worse, the exit time is very long. Does java need to write to the disk on exit?
OK. I guess what I'm trying to do is use an IDE (Eclipse) which runs on Java, to compile apps for java. I've written some small apps using a text editor and compiling the source from the dos cmd line. Seems resonable. But as soon as I get into one of these IDEs the system requirements goes into the gigabytes. Is there an option for a lightweight dev environment?
-
We are working on AIX 6.1 with Oracle 11gR2, we are facing any issue during the restore operation and the RMAN is not existing after restore
channel ORA_SBT_TAPE_2: reading from backup piece sun2gm3b_1_1
channel ORA_SBT_TAPE_1: piece handle=svn2g23b_1_1 tag=TAG20120204T220011
channel ORA_SBT_TAPE_1: restored backup piece 1
There is no contention from the MML (IBM TSM), there is no CPU contention on the Oracle Database Server, no errors in the alert log file and through RMAN backup pieces have also been checked using the RMAN VALID command.
Please any hints or indicator?
Thank you very much.user8676543 wrote:
We are working on AIX 6.1 with Oracle 11gR2, we are facing any issue during the restore operation and the RMAN is not existing after restore
channel ORA_SBT_TAPE_2: reading from backup piece sun2gm3b_1_1
channel ORA_SBT_TAPE_1: piece handle=svn2g23b_1_1 tag=TAG20120204T220011
channel ORA_SBT_TAPE_1: restored backup piece 1
There is no contention from the MML (IBM TSM), there is no CPU contention on the Oracle Database Server, no errors in the alert log file and through RMAN backup pieces have also been checked using the RMAN VALID command.
Please any hints or indicator?
RMAN> EXIT -
Running RMAN 10.2.0.1 on AIX 5.3
All backups are successful.
Exit is successful when connected to other(non-RMAN) databases.
The problem occurs when I am backing up the RMAN catalog.
Once connected to the target database (in this case the rman catalog), the exit command hangs.
I can reproduce this whether connected to the backup rcvcat or not.
This happens on both the primary catalog and the backup catalog.
Thank you,
Randy.Why should you backup the RMAN catalog in the same RMAN database. Thats pointless. Use other backup means for backing catalog database. That is what i would suggest. Use cold backup or tapes to backup your RMAN database.
-
RMAN FULL-INCR ARCHIVE SCRIPT ON WINDOWS
HI Guys
schedule of RMAN full/incr and archive backup in windows environmental i hv to schedule through scheduler.
request to please send me script which i can schedule.
ThankWrite a batch script. And schedule it with windows task scheduler.
For example
contents of rman.bat
SETLOCAL
set ORALOG=D:\RMAN_BACKUP
set SCR_HOME=D:\oracle\bin
set DB_NAME=DBNAME
set CMDFILE=%SCR_HOME%\rman_script.txt
set LOGFILE=%ORALOG%\rman_log_file.log
set MAILFILE=Message.log
del /q %LOGFILE%
%SCR_HOME%\rman target / nocatalog cmdfile=%CMDFILE% log=%LOGFILE%
findstr /I "ERROR" %LOGFILE%
if %errorlevel% equ 1 goto noerrorfound
goto errorfound
:errorfound
echo Database : %DB_NAME% > %MAILFILE%
echo Hostname : %COMPUTERNAME% >> %MAILFILE%
echo Date : %DATE% >> %MAILFILE%
echo Time : %TIME% >> %MAILFILE%
echo RMAN Log File : %LOGFILE% >> %MAILFILE%
echo Log File İçeriği: >> %MAILFILE%
type %LOGFILE% >> %MAILFILE%
cscript mail.vbs "FAILED RMAN Backup" %MAILFILE%
exit
:noerrorfound
echo Database : %DB_NAME% > %MAILFILE%
echo Hostname : %COMPUTERNAME% >> %MAILFILE%
echo Date : %DATE% >> %MAILFILE%
echo Time : %TIME% >> %MAILFILE%
echo RMAN Log File : %LOGFILE% >> %MAILFILE%
echo Log File icerigi: >> %MAILFILE%
type %LOGFILE% >> %MAILFILE%
cscript mail.vbs "SUCCESSFUL RMAN backup" %MAILFILE%
exit
contents of rman_script.txt
--your backup script for example
backup database; -
My issues with iPhone 2.0 on Original iphone
Here are some of the issues i am seeing with the 2.0 update on my iPhone 2g.
Texting is sometimes slow, however, exiting the app, or rebooting fix's the issue.
Even though Mail was properly configured, upon the first install of 2.0, mail would crash each time I would try to open the app. To fix this, I re-installed 2.0.
Fewer bars then in 1.1.4, in my house I normally get 4 bars, but now I am lucky to get 2.
Slower system response when playing MP3's. I was playing some music today and I wanted to send a SMS, so when I tried to open the app, the screen showed a bit of the app with a white background, but no text. About 2 or 3 seconds later, it pulled in the data.
Is anyone else having these issues and/or are they being addressed by Apple?
KurtI have this issue too - as do others posting to macrumors forum.
Apple - What's ging on with Updating Files that haven't changed? Mine does this to ~400 - 500 tracks every time i sync. -
We are using an external WMS system (PKMS), so SAP is configured with
de-centralized warehousing. This means that when product is shipped we
receive from the external WMS an iDoc for the Post Goods Issue transaction.
They come in bunches, since there can be many orders on a truck. It seems
that the processing does not occur quick enough and some iDocs are blocked
because the SKU they are trying to update is locked while being updated. We
have turned on the option of 20 retries but still get too many iDocs
blocked on lock. The lock is always from the RFC user id, so the interface
is blocking itself.Many inbound SHP_OBDLV_CONFIRM_DECENTRAL idocs coming in from SBC (from |
an external WMS system) are being rejected with the error message M3024
Valuation data for material nnnnn is locked by the user RFCCI01.
For business reasons, these idocs are processed immediately and
therefore come into the system in a batch. Even with low number of
idocs coming in, up to 50% of them return with a status 51. For
example, of 13 idocs coming in 7 are going to status 51 with this
error.When we reprocess them with program RBDMANIN this obviously
reprocesses
them properly because it processes them sequentially preventing locking
|however this does not meet our business need.Hi Scott!
In case you are getting a rejection of up to 50% of the IDOCs, then you should go for sequential processing as a immediate help (no parallel processing, not the 'real' sequential technic).
This might look slower in first hand - but with such high rejection rates you are anyway not faster currently. And it's much better to have more successful IDOCs than faster errors.
In a second step you should look for performance optimizations. If you are using Retail: there is an option for a late article master lock option (SPRO general data). It has also business effects, but should be discussed.
Also slow user-exits in update tasks can be a reason for long locks during booking -> check coding for slow parts.
Regards,
Christian -
Oracle mentions Red Hat, Suse, and TurboLinux in its tech docs. Has anyone used more than one of these and done any performance tests. How about general reliability comparisons? We are also interested in how well their support depts handle real problems.
Thanks.976963 wrote:
No, It is not only about RMAN Slow performance, But that's what you are describing as your issue.
Please advice me your suggestions to improve RMAN backup performance I have no clue as to how to investigate RMAN functionality. I've never use the software nor expect ever to use it. However I do recognize OS and Hardware issues and that is why I'm the primary moderator in those OS and Hardware forums.
By the way, you have a simultaneous post already in the Linux forum
Slow IO Performance in Oracle uek 64bit kernel
which might be a duplicate representation of this same issue.
For the future, consider that such multi-posting is frowned upon.
Leave this thread here.
Use that thread to investigate OS issues and only OS issues. -
Query_Data_Source_Name question
Dear all,
In my form, there is a block 'B_BLOCK' used 'from clause query' method to retrieve data. i.e. Set_Block_Property ('B_BLOCK', Query_Data_Source_Name, v_sql) will be used. In this multi records block, I set the Query Arrary Size = 0, Number of Records Buffered = 0, Number of records displayed = 4, Query all records = No, Maximum Query Time = 0, Maximum Records Fetched = 0.
My question is, e.g. there should be 8000 records after execute query. But the screen can only display 4 records at the same time. User need scroll down to view the records. I wonder is there any method that I know how many records are already fetched. That is, although there are 8000 records, user maybe only scroll to records 150 and then scroll up to the first record. But I want to know the total fetched record is 150.
I need to know no. of records are already fetched because when user pressed Exit button, I have a checking before exit the form. There is a loop for this block to check any mismatch data. If there are 8000 records, I believe although user only read 150 records, the loop will go through all the records and therefore the speed is very slow to exit the form. If I know the number of records already fetched, then, I can loop only 150 records.
Hope there is solution to solve the problem. Thanks for any advise and thanks in advance.
Regards.I think this is so because Forms locks the table opening the Form.
I couldn't find possibility to unlock the table.
But you can the following. Instead to base the block on a table you can resd in the trigger "WHEN NEW BLOCK" the rows and insert them in the block using the NEXT_RECORD buil-in:
NEXT_RECORD built-in:
Navigates to the first enabled and navigable item in the record with the next higher sequence number than the current record. If there is no such record, Form Builder will fetch or create a record. If the current record is a new record, NEXT_RECORD fails.
Using the record status it is possible to insert, update, delete the rows in the DB table. -
Hello Everyone,
Social Apps on my Asha 311 is not working properly, need help!!
*#0000#
Nokia 311
V 05.92
10-10-12
RM-714
(c) Nokia
Variant 00.00
Notifications App - V1.5.13
Line - V1.5.2
Wechat - V2.0.2
Viber - 2.22.25
FB - 3.2.2
(app source: Nokia Store)
When i open any of these apps, it works fine in the first few seconds, then it closes and the phone goes into a loading screen that cannot be closed. Already did a full factory reset (erase all data) and downloaded the apps again. What should i do??? if i open the notifications app, i cant close it. I have to turn off the phone then turn it on again. Thank You in advance!!May it's About "RAM"
and How many Apps You Installed On Ur phone!
cuz I have Asha 501 and When I Install many Apps it Works Slow and Exit Sometimes!
IamAJBLove -
Aperture 3: Many new problems...
Hi,
I recently upgraded to Aperture 3.0, then subsequently to 3.0.1. I am noting the following issues:
1. Thumbnail previews of photos are sometimes distorted and horizontally stretched (especially photos that are in a vertical orientation)- they also load VERY, VERY slowly.
2. Import of photos from a Nikon D300 is super-slow, way slower than when Aperture 2 was being used
3. Frequent and unexplained crashes (program shuts down for no reason).
4. Overall, the program runs much slower than the previous version (slower to load, slower to view photos, slower to exit).
I have 10 GB of RAM on an 8-core Mac Pro- I highly doubt my computer can't handle this program. Nothing else has changed on my computer. I have performed all the latest OS and Aperture updates (ie running 10.6.2 and Aperture 3.0.1).
None of this occurred with the previous version. Why is this occurring now?
Is it me, or this program now a step back compared to the previous version?
Most problematic are the problems with the thumbnail previews. Someone please help! Thanks.Sorry pvonk,
There is something serious wrong with A3 or A3.0.1 Why do I say that? Because of a very simple experiment I did
1. Virgin mac, no previous exposure to Aperture.
2. C2D with 2 GB RAM with minimal third party programs
3. Install A3
4. Lots of crashes from an imported library
5. Upgrade A3.0.1
6. Create a new Library on internal drive
7. Import 418 D700 NEF files into this virgin library (NO A2 library IMPORT), all referenced on external drive
8. Did a global change on all the pictures by
a. Correcting white balance from grey card image
b. Adjust contrast to 0.15
c. Adjust vibrancy to 0.1
d. Copy settings and stamp to about 1/2 the library
e. Repeated steps a-d on the second half of the library
9. Zoom to 100% to begin editing and waited 134 seconds for the spinning beach ball to go away
10. Open retouch tool and waited 8 minutes for the spinning beach ball to go away... finally did a forced quit
11. Report sent to Apple
12. Reopen A3, return to same picture and do 100% zoom
13. 3-4 retouches and A3 crashes
14. Report sent to Apple
15.Reopen A3 and return to same picture
16. Did some retouching, dodging, and retouch and crash.
17. Report sent to Apple.
The library was built entirely off of original RAW camera files
So I purchased a complete copy for my test machine and an upgrade copy for my two working machines. I installed the full copy on a clean machine to make sure it was ok and I understood it before I installed in on my working machines. My simple experience is described above.
BOTTOM LINE: A3 exhibited multiply crashes and hangs way beyond expectations. Even on the very simple test described above, it seriously failed. I as much as I want to start using it to process my professional work (35-40000 images a year), it really is not ready for prime time. Something is seriously wrong with it. I returned my unopened upgrade package to the store and got a refund. The Apple person said something interesting to me. Well no one else has returned it, but have you ever tried to return opened software. No way. The only way I was able to return the upgrades was I purchased a full version and decide its performance was subpar and then I returned the UNOPENED upgrades.
CONCLUSION: I will continue to test A3 as new upgrade become available on my full version but since it only takes me 30 minutes to cause multiple failures, this is not a huge investment in my time. When A3 becomes stable on my test bed, with enormous care and caution I will upgrade one of my working machines and make sure it is as stable as A2.1.4
As I have said before, if I need the top end of a machine (Quad with tons of memory) for A3 to work, a much more economical solution will be to switch to LR. This is a pretty important decision for me since I have more than 100,000 images in Aperture structures and would rather not switch, but the current A3 does not achieve acceptable performance. The concepts and features are absolutely wonderful. The implementation is perhaps the worse I have ever seen. I will optimistically wait for A3.0.2. -
A workaround was suggested by a member of the community to turn off both hardware and Flash acceleration. It worked fine (no crashes since), but runs very slowly. In particular, takes five seconds to exit and is often still listed as a running process. Very slow in connecting to Web pages, and very slow loading them because of the graphics. Very slow in loading video. I expected slower responses, but this is REALLY slow. I'm running 64-bit Windows 7 and an Nvidia GE Force 7800 graphics card with all the drivers updated and the plugins for Firefox mostly set on "ask to activate". Should I expect this much reduction in performance when the workaround I mentioned was put into place? If so, it's half a loaf at best. The only thing questionable is that I have two Youtube downloaders that I am trying, but I made the assumption that these were only applied when you downloaded something from Youtube.
In case you are using "Clear history when Firefox closes": do not clear the Cookies
If you clear cookies then Firefox will also try to remove cookies created by plugins and that requires to start plugin-container processes that can slow down closing Firefox.
Instead let the cookies expire when Firefox is closed to make them session cookies.
*Firefox/Tools > Options > Privacy > "Use custom settings for history" > Cookies: Keep until: "I close Firefox"
Maybe you are looking for
-
Propagation of Credentials from ADF Pages to BPEL
It is easy and straightforward to call web services from ADF pages. But I am a bit confused about the security issues around use of BPEL and OWSM with ADF. Suppose that I have authenticated against an LDAP to access an ADF page and the page make call
-
Does anyone have any advice on which DVD ripper/converter to iPod friendly format (commercial or free, though free is obviously better for me) for windows platform cheers
-
How to access SDM gui in windows
hi, i want to deploy/undeploy something using SDM (windows installation), what is the path for the SDM gui,.,....
-
in the first itunes i had the option show dubbel parts but whit the newer one i cant find that option annymore, where is it?
-
Direct production in collective order network, automatic 261...
hi all, i have a problem in a schenario with collective order network. i have order B part of collective order network of order A. and material of order B has special procurement key 52 maintained in it. and colletive order of goods movement is selec