Rolling Upgrades in Rac

Hi All,
can you please explain abt rolling upgrades in rac env
suppose i am having rac setup on 10.2.0.1 if i want to apply patch set 10.2.0.4 is it required downtime
As per the readme of 10.2.0.4
1)Stop every thing on node 1 and apply patch set
2)stop every thing on node 2 and apply patch set
3)stop the database and run the catupgrd.sql script.
My question is for running catupgrd script i have to bring down the db so what the meaning of rolling upgrade??
And in this type of sece which type of upgrade is best one rolling or non-rolling.
KK
Edited by: user12061936 on May 20, 2010 2:35 PM

Hi HTH
Thank you for clearing my doubts regarding Rolling upgrades and i have gone though the link and i have noticed the following steps
For Standalone system
1) Install Oracle 10g to a separate Oracle Home on the same server
2) Create a 10g database with only the base tablespaces: SYSTEM, SYSAUX, UNDO, and TEMP
3) On the 9i database, put all your tablespaces into read only mode (write downtime begins)
4) Perform a transportable tablespace export of all non-system tablespaces (as a sysdba user)
5) Shut down the 9i database (true downtime begins)
6) Start up the 10g database
7) Perform a transportable tablespace import into the 10g database (end true downtime)
8) Make all your tablespaces read/write (end write downtime)
The above steps is going to work only for 9i to 10g or can we use it to any version to version.
BY using TTS is it suggestable method to do the upgrade??
KK
Edited by: user12061936 on May 21, 2010 11:41 AM

Similar Messages

Rolling upgrade in MAA database with 2 node RAC on each site

All,
I have 2 node rac on SITE A, and it has Physical Standby database on SITE B it also 2 node RAC. Currently it is on 10.2.0.4 on AIX systems. We have mission crictical application runing and we can't aford the downtime of 5 to 6 hr to upgrade our CRS,ASM,DB oracle homes to 11G or even 10.2.0.5. I read rolling upgrade document on metalink and also documnet as well, and come up with following sequence.
SITE A==Primary Site, SITE B == Standby database
1) Shutdown DB on SITE B
2) Take the OS Level Backup of Binaries only on SITE B (Both node), I think this is just a Precautionary Measure
3) Start the DB on SITE B Still Physical Standby database let it synch with primary Site SITE A
4) Convert SITE B to Logical standby database
5) let it run for some time and be the two site SYNCH
6) now stop archivelog shipping to STANDBY SITE SITE B
7) Shutdown SITE B and apply the patch/upgrade binaries (CRS/ASM/DB)
8) Start and upgrade the DB on SITE B
9) Start the logical standby apply (SQL APPLY) and let it catch up with SITE A
10) Now Switchover to logical standby database on SITE B and ask application to connect to SITE B
11) Make sure the archivelogs are moving from SITE B to SITE A and SQL Apply is runing.
12) Now Stop archive log flow from SITE B to SITE A
13) Stop the database/ASM/CRS on SITE A
14) upgrade the SITE A binaries (CRS/ASM/DB)
15) Start and upgrade the DB on SITE A
16) Start the logical standby apply (SQL APPLY) and let it catch up with SITE B
17) Now Switchover Back to SITE A and let the application connect to SITE A
At this point by SITE B is a Logical Standby database
I think these summaries the steps involved in rolling upgrade, but I can't understand how I can go back to my MAA with physical standby database mode?
Does anyone has done this? Can anyone give me some tips or comments on my steps??
Help me if you can... thanks in advance.

Hi,
Is it you want to know the OS/Kernel side setting? You must comply with the following kernel requirements to install the clusterware.
http://docs.oracle.com/cd/B28359_01/install.111/b28259/pre_hpux.htm#BABJHCJI
Thanks.

Rolling Upgrade of Oracle 10gR2 RAC with Physical Standby (DG)

Hi DBAs,
I want to do a rolling upgrade of Oracle 10.2.0.3 (2 node RAC+ASM) with Physical Standy Dataguard to Oracle 10.2.0.4 also applying CPU (April2009).
Please suggest the best way to perform this upgrade and patching with no downtime.
Thanks
-Samar-

Unfortunately rolling upgrade is only possible with a logical standby in place:
http://download.oracle.com/docs/cd/B19306_01/server.102/b14239/upgrades.htm#CHDBJAGG
That doesn't help you at this time, but 11g has a way to do it with a physical standby. Even here you need an intermediate logical standby:
http://download.oracle.com/docs/cd/B28359_01/server.111/b28294/rollup.htm#CHDHCBGA
Werner

RAC rolling upgrade

Hi all
I'm performing rolling upgrade of cpu patch at rac environment
I open a connection to the database (using service) and it connects to the first node. Then I shutdown the first node to perform an upgrade and check the connection, but it doesn't switch to the second node automatically, but returns TNS error. What about if it's a production database and there're a lot of online scripts running on that node?
However, I tried it again and it automatically switched wihout any error. What can be reason?
thanks
John

Agreed... use the -local option and do each node individually.
11gR2? If you have TNS issues, it is indicative of an improperly configured local/remote_listener or SCAN on one or more nodes. Sometimes when additional nodes are added, these parameters may not necessarily be correct.

Upgrade of RAC database in 10g or 11g

Hi,
Can we upgrade RAC database without downtime like we do patch?
Regards

user602441 wrote:
Hi,
Can we upgrade RAC database without downtime like we do patch?
RegardsHi,
The Oracle Clusterware software always fully supports rolling upgrades, while the ASM software is rolling upgradeable at version 11.1.0.6 and beyond.
Rolling upgrade: we mean upgrading software (Oracle Database, Oracle Clusterware, ASM or the OS itself) while the cluster is operational by shutting down a node, upgrading the software on that node, and then reintegrating it into the cluster, and so forth one node at a time until all the nodes in the cluster are at the new software level.
For the Oracle Database software (RAC), it is possible only for certain single patches that are marked as rolling upgrade compatible. Most Bundle patches and Critical Patch Updates (CPU) are rolling upgradeable. Patchsets and DB version (10g to 11g) changes are not supported in a rolling fashion, one reason that this may be impossible is that across major releases, there may be incompatible versions of the system tablespace, for example. To upgrade these in a rolling fashion one will need to use a logical standby with Oracle Database 10g or 11g.
Regards,
Levi Pereira

Database rolling upgrade cpu patch with minimal service outage

Oracle 11gR2 Primary Instance
Oracle 11gR2 Active Physical Standby
Requirement of applying CPU and CVE patches with minimal downtime. I believe that I've been able to apply CPU patch to Active Standby after deferring redo transport and applying patch to active physical standby first and then performing a manual switchover to have the standby become the primary and former-primary become the standby and they apply CPU Patch there and then perform another role transition to its original role.
Appreciate your experience with primary and active physical standby rolling upgrades/patches. Thanks

Hello;
I believe you are over thinking this. All you have to do on the Standby is install the patch. No switchover, no SQL. No nothing.
The SQL will move from the Primary to the Standby.
Check the readme for the patch closely or post the patch number if you want to confirm.
OR
How do you apply a Patchset,PSU or CPU in a Data Guard Physical Standby configuration [ID 278641.1]
( If it helps I had the same thoughts after my first Data Guard setup !! )
If you are trying to avoid downtime entirely the switchover will take about as long as the patch.
If you have RAC this is interesting :
http://javeedkaleem.blogspot.com/2010/05/how-to-apply-oracle-database-patch.html
As is this :
http://docs.oracle.com/cd/B28359_01/server.111/b28282/schedule_outage.htm
Best Regards
mseberg

Oracle physru rolling upgrade problem

Hi, I'm having a problem with the Oracle Physru script provided from MOS note # 949322 I was hoping I could get some help with.
My system contains of two Oracle 11.1.0.7.0, one primary and one physical standby.
The hosts are two Oracle Solaris 10 Sparc x64 machines.
My goal is to update from 11.1.0.7.0 to 11.1.0.7.8 using the script that is provided, however I'm bumping into problems at the step where the script is checking the Apply lag on the logical standby (function "get_apply_lag"). The lag seems to increase, and to be that indicates a problem with the redo apply process. However, when I query the DBA_LOGSTDBY_EVENTS View i get the following:
SQL> SELECT EVENT_TIME, STATUS, EVENT FROM DBA_LOGSTDBY_EVENTS ORDER BY EVENT_TIMESTAMP, COMMIT_SCN;
EVENT_TIME STATUS EVENT
20-SEP-11 16:02:12 ORA-16111: log mining and apply setting up
20-SEP-11 16:02:12 Apply LWM 848622, HWM 848622, SCN 848622
Showing the output from Primary archive dest 2:
SQL> show parameter log_archive_dest_2;
NAME TYPE VALUE
log_archive_dest_2 string service="test08db2", LGWR ASYNC NOAFFIRM delay=0 OPTIONAL compression=DISABLE max_failure=0 max_connections=1 reopen=300 db_unique_name="db_test08db2" net_timeout=30 valid_for=(online_logfile,primary_role)
SQL> select from v$LOGSTDBY_STATS;*
NAME VALUE
logminer session id 1
number of preparers 1
number of appliers 11
server processes in use 15
maximum SGA for LCR cache (MB) 50
maximum events recorded 2000000000
preserve commit order TRUE
transaction consistency FULL
record skipped errors Y
record skipped DDLs Y
record applied DDLs N
NAME VALUE
record unsupported operations Y
realtime apply Y
apply delay (minutes) 0
coordinator state IDLE
coordinator startup time 20-SEP-11 16:02:11
coordinator uptime (seconds) 727
txns received from logminer 62
txns assigned to apply 29
txns applied 29
txns discarded during restart 33
large txns waiting to be assigned 0
NAME VALUE
rolled back txns mined 4
DDL txns mined 2
CTAS txns mined 0
bytes of redo mined 8195884
bytes paged out 0
pageout time (seconds) 0
bytes checkpointed 709802
checkpoint time (seconds) 0
system idle time (seconds) 479
standby redo logs mined 0
archived logs mined 4
NAME VALUE
gap fetched logs mined 2
standby redo log reuse detected 0
logfile open failures 0
current logfile wait (seconds) 0
total logfile wait (seconds) 0
thread enable mined 0
thread disable mined 0
distinct txns in queue 0
41 rows selected.
SQL> select type, high_scn, status, pid from v$logstdby order by type;
TYPE HIGH_SCN STATUS PID
ANALYZER 850320 ORA-16116: no work available 20702
APPLIER 848808 ORA-16116: no work available 20719
APPLIER 850320 ORA-16116: no work available 20731
APPLIER 850204 ORA-16116: no work available 20727
APPLIER 848895 ORA-16116: no work available 20725
APPLIER 848665 ORA-16116: no work available 20705
APPLIER 848677 ORA-16116: no work available 20709
APPLIER 848728 ORA-16116: no work available 20713
APPLIER 848740 ORA-16116: no work available 20715
APPLIER 848796 ORA-16116: no work available 20717
APPLIER 848842 ORA-16116: no work available 20721
TYPE HIGH_SCN STATUS PID
APPLIER 848854 ORA-16116: no work available 20723
BUILDER 850320 ORA-16116: no work available 20221
COORDINATOR 850326 ORA-16116: no work available 20119
PREPARER 850318 ORA-16116: no work available 20223
READER 850326 ORA-16116: no work available 20217
16 rows selected.
Physru Script Output:_
### Initialize script to either start over or resume execution
Sep 20 14:35:41 2011 [0-1] Identifying rdbms software version
Sep 20 14:35:41 2011 [0-1] database nobilldb is at version 11.1.0.7.0
Sep 20 14:35:42 2011 [0-1] database nobilldb is at version 11.1.0.7.0
Sep 20 14:35:44 2011 [0-1] verifying flashback database is enabled at db_test08db1 and db_test08db2
Sep 20 14:35:44 2011 [0-1] verifying available flashback restore points
Sep 20 14:35:45 2011 [0-1] verifying DG Broker is disabled
Sep 20 14:35:46 2011 [0-1] looking up prior execution history
Sep 20 14:35:46 2011 [0-1] purging script execution state from database db_test08db1
Sep 20 14:35:46 2011 [0-1] purging script execution state from database db_test08db2
Sep 20 14:35:47 2011 [0-1] starting new execution of script
### Stage 1: Backup user environment in case rolling upgrade is aborted
Sep 20 14:35:47 2011 [1-1] stopping media recovery on db_test08db2
Sep 20 14:35:48 2011 [1-1] creating restore point PRUP_0000_0001 on database db_test08db2
Sep 20 14:35:49 2011 [1-1] backing up current control file on db_test08db2
Sep 20 14:35:50 2011 [1-1] created backup control file /opt/oracle/product/11.1.0.7/dbs/PRUP_0001_db_test08db2_f.f
Sep 20 14:35:50 2011 [1-1] creating restore point PRUP_0000_0001 on database db_test08db1
Sep 20 14:35:51 2011 [1-1] backing up current control file on db_test08db1
Sep 20 14:35:52 2011 [1-1] created backup control file /opt/oracle/product/11.1.0.7/dbs/PRUP_0001_db_test08db1_f.f
NOTE: Restore point PRUP_0000_0001 and backup control file PRUP_0001_db_test08db2_f.f
can be used to restore db_test08db2 back to its original state as a
physical standby, in case the rolling upgrade operation needs to be aborted
prior to the first switchover done in Stage 4.
### Stage 2: Create transient logical standby from existing physical standby
Sep 20 14:35:53 2011 [2-1] verifying RAC is disabled at db_test08db2
Sep 20 14:35:53 2011 [2-1] verifying database roles
Sep 20 14:35:54 2011 [2-1] verifying physical standby is mounted
Sep 20 14:35:54 2011 [2-1] verifying database protection mode
Sep 20 14:35:55 2011 [2-1] verifying transient logical standby datatype support
Sep 20 14:36:00 2011 [2-2] starting media recovery on db_test08db2
Sep 20 14:36:11 2011 [2-2] confirming media recovery is running
Sep 20 14:36:12 2011 [2-2] waiting for v$dataguard_stats view to initialize
Sep 20 14:36:13 2011 [2-2] waiting for apply lag on db_test08db2 to fall below 30 seconds
Sep 20 14:36:13 2011 [2-2] apply lag is now less than 30 seconds
Sep 20 14:36:14 2011 [2-2] stopping media recovery on db_test08db2
Sep 20 14:36:15 2011 [2-2] executing dbms_logstdby.build on database db_test08db1
Sep 20 14:36:27 2011 [2-2] converting physical standby into transient logical standby
Sep 20 14:36:52 2011 [2-3] opening database db_test08db2
Sep 20 14:37:28 2011 [2-4] configuring transient logical standby parameters for rolling upgrade
Sep 20 14:37:29 2011 [2-4] starting logical standby on database db_test08db2
Sep 20 14:37:37 2011 [2-4] waiting until logminer dictionary has fully loaded
Sep 20 14:37:58 2011 [2-4] dictionary load 17% complete
Sep 20 14:38:09 2011 [2-4] dictionary load 32% complete
Sep 20 14:38:19 2011 [2-4] dictionary load 43% complete
Sep 20 14:38:30 2011 [2-4] dictionary load 59% complete
Sep 20 14:38:40 2011 [2-4] dictionary load 62% complete
Sep 20 14:38:50 2011 [2-4] dictionary load 70% complete
Sep 20 14:39:01 2011 [2-4] dictionary load 72% complete
Sep 20 14:39:11 2011 [2-4] dictionary load 75% complete
Sep 20 14:40:54 2011 [2-4] dictionary load is complete
Sep 20 14:41:00 2011 [2-4] waiting for v$dataguard_stats view to initialize
Sep 20 14:41:01 2011 [2-4] waiting for apply lag on db_test08db2 to fall below 30 seconds
Sep 20 14:42:02 2011 [2-4] current apply lag: 316
Sep 20 14:42:32 2011 [2-4] current apply lag: 316
Sep 20 14:43:03 2011 [2-4] current apply lag: 376
Sep 20 14:43:33 2011 [2-4] current apply lag: 376
Sep 20 14:44:03 2011 [2-4] current apply lag: 437
Sep 20 14:44:34 2011 [2-4] current apply lag: 437
Sep 20 14:45:04 2011 [2-4] current apply lag: 497
Sep 20 14:45:35 2011 [2-4] current apply lag: 497
Sep 20 14:46:05 2011 [2-4] current apply lag: 558
Sep 20 14:46:36 2011 [2-4] current apply lag: 558
Sep 20 14:47:06 2011 [2-4] current apply lag: 618
I would appreciate any help I could get, I'm stuck =/
Regards,

http://www.oracle.com/technetwork/database/features/availability/maa-wp-11g-upgrades-made-easy-131972.pdf

Rolling upgrade

Hi all,
How to know if my Oracle home is shared to install the patcher 10.2.0.3 to 10.2.0.4 in rolling upgrade for Oracle Clusterware?
Which refers to the shared homes, only CRS Home or Oracle Home and CRS Home?
I have 3 nodes on RAC.
SO: Solaris Operating System (SPARC 64-Bit)
RDBMS: Oracle 10g R2
Amin

A shared Oracle home is one where a single Oracle Home is accessed via shared storage from multiple nodes. Contrast this with the case needed for rolling upgrade, where there would be a separate Oracle Home installed on each node in the RAC.
The CRS_HOME and ORACLE_HOME should be different in any event (they can be under the same Oracle Base).
Does that help?

Rolling upgrade from 11.2.0.2 to 11.2.0.3

Windows 2008. 2 node RAC
I attempted a rolling upgrade from 11.2.0.2 to 11.2.0.3 in a dev environment in prep to do my prod. I hit all kinds of issues and when googling I found there was a known issue with rolling upgrades between these versions and there was a patch I was meant to have applied before starting the rolling upgrade. I rolled back to 11.2.0.2 but for the life of me, I cant find reference to that patch now.
anyone know what it is?
/still looking/

Hello,
Please refer below document
Things to Consider Before Upgrading to 11.2.0.3 Grid Infrastructure/ASM (Doc ID 1363369.1)
You can check the section related to 11.2.0.2.
It is better to check complete document carefully before attempting the upgrade.
Regards,
Vijay

OS upgradation For RAC

Hi guys...How is everything.....
I have Node RAC on Windows2003R2 Itaniam. I Need to know is there any procedure for OS upgradation for RAC..I mean shall I down the Cluster & Database when OS up gradation..

Hi JIC,
Please be more specific. Do you plan upgrade OS for Win2008? Or did you means service packs and patches through windows update?
In general, if your OS is certified by Oracle to run RAC if you are using service pack 1,2,3 doesn't matter, they are certified too.
"The Oracle Clusterware and Oracle Real Application Clusters both support rolling upgrades of the OS software when the version of the Oracle Database is certified on both releases of the OS (and the OS is the same, no Linux and Windows or AIX and Solaris, or 32 and 64 bit etc.). This can apply a patch to the operating system, a patchset (such as EL4u4 to EL4u6) or a release (EL4 to EL5). Stay within a 24 hours of upgrade window and fully test this path as it's not possible for Oracle to test all these different paths and combinations."
This rule is valid to Unix/Linux and windows Systems.
You can check it on My Oracle Support note ID 220970.1
Best Regards,
Rodrigo Mufalani
www.mufalani.com

Is it possible to delay a rolling upgrade after a single domain is updated?

Hello,
I am working on new Azure services for my group. For various reasons, we have decided that a Staging / VIP-Swap deployment is not useful for our group and will be going with a rolling upgrade. What I am interested in, is knowing if there is a way to
"pause" a rolling upgrade after the first update domain until it is released to complete. Or, for a set time ( 30 minutes, etc ), to allow us the opportunity to essentially have a A-B deployment of our webrole, to observe behavior while receiving
production traffic before the release has completely deployed.
Thanks for your help
Jeff

hello Jeff,
Currently, base on my experience, delay a rolling upgrade or "pause" a rolling upgrade after the first update domain until it is released to complete, it may be difficult. I suggest you could can submit a feature suggestion on
http://www.mygreatwindowsazureidea.com/forums/34192-windows-azure-feature-voting. I suggest you could get more detail about upgrade domain and fault domain (http://blog.toddysm.com/2010/04/upgrade-domains-and-fault-domains-in-windows-azure.html).
Best Regards,
Will
We are trying to better understand customer views on social support experience, so your participation in this interview project would be greatly appreciated if you have time. Thanks for helping make community forums a great place.
Click
HERE to participate the survey.

Upgrade oracle rac 10g r2 10.2.0.3 to 10.2.0.4

Hi all,
I goint to upgrade oracle rac 10g r2 10.2.0.3 to 10.2.0.4 on Solaris 9, in the README file(patchset) does not say that upgrade the OCR, but int he Doc ID 316889.1 to say; it is necessary to upgrade or nothing happens if Iit does not upgrade?
Mexman
Edited by: ACS on 28/10/2009 12:19 PM

Hi!
I suppose if you have a Metalink note on which explains is necessary update OCR you must follow it. Software maker always are in reason!
Chiao!

Scshutdown fails - "Rolling Upgrade in progress"

Anybody seen this before?
bash$ sudo scshutdown -g 0 -y
scshutdown: Cannot shutdown the cluster. Rolling Upgrade in progress.
bash$
No one has (with intention) started an upgrade on this cluster.. Its SC3.1 U2 on Solaris 8.

While you are saying that no one hast started an upgrade, is it possible that this cluster got recently patched with current Sun Cluster patches?
If so, you might want to check if the follwoing binary exists on your cluster:
/usr/cluster/bin/scversions
If it exists, just call it.
If you see something like:
# scversions
Upgrade commit is NOT needed. All versions match.
then I don't know what your problem is. However, if it says something different, you might want to try:
# scversions -c
and see if the problem went away.
Greets
Thorsten

Upgrade oracle RAC 10.2.0.1 to 10.2.0.4 software only without database

I want to upgrade oracle RAC 10.2.0.1 to 10.2.0.4 software only without database.
In this environment i am using OCFS file system not asm . please send me step to configure.

In other terms its called patchset application into your database home. Go to this note, assuming its linux, and do all steps until step 9 which is database upgrade.
10.2.0.4 Patch Set - Linux x86 64bit README [ID 559255.1]
Regards
Ed

Rolling upgrade from 11.1.0.6 to 11.1.0.7

Hello -
I am testing the "rolling upgrade" from 11.1.0.6. to 11.1.0.7. I have a two node cluster on Linux. Is there a way to upgrade either the ASM software or the dbms software without causing downtime to the entire cluster? I can upgrade the clusterware in a true rolling fashion. However, when I attempt to upgrade the ASM location, it ends up taking down the other node. Same for the rdbms software - The other instance ends up going down when I start the patching on one instance - from the remove operations being performed. Has anyone successfully patched an entire stack (without downtime)?
Thanks in advance for any insights!

Yes, I am following those instructions. The ASM instructions caused the other node to be inaccessible. But, when I did the rdbms upgrade, then node B was also inaccessible during that time. I will go through more iterations of this testing, but so far, I have not been able to develop a proof of concept for a true rolling upgrade from 11.1.0.6 to 11.1.0.7. If anyone has been able to do this, please let me know the procedures that were followed.
Thanks,
Mike

Rolling Upgrades in Rac

Similar Messages

Maybe you are looking for