on 09-17-2007 5:46 AM
Hi all,
We wanted to do a cluster failover on our Production system.
Its an ECC 5.0 system with Oracle 10G on AIX
The scenario is
During the cluster failover we need to unmount all the file systems from our production and mount it on the Integration server and make the Production up.The integration system FS will be unmounted.
This scenario works good but we are having problem with the SAP profiles
The production system is on hostname1, and integration is on hostname2
In production it uses the profile SID_D*_hostname1 and when we do the cluster failover,it starts taking the profile of hostname2.
Though we have an alias name for the system and the profiles also with the alias names,we are unable to bring up the production on these profiles in either of the hosts.
Regards
Bharathwaj V
The problem is solved.
When the cluster failover is happening,dbstartup failed as the bin directory path was set to old oralcle home 920_64.When this was set to 102_64 the SAP started with out any problem.
Bharathwaj V
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello Bharathwaj,
Thanks for letting us know, it is good that you found out what was wrong. Sorry I never replied to your last questions, I haven't posted anything the last 3 weeks as I became a father of a little baby girl on the 21st of September ! She does keep me busy!
All the best
Andreas
Hello,
I would follow a phased approach on this
First try to see if you can bring up the production database only. Make sure that the listener is started first. Check the orasid user environment variables and see if they are set to what you would expect them to be
After that, try to start the database from orasid with sqlplus "/ as sysdba" and startup
Secondly, make sure that the sidadm environment is right in the new server. If you sent here what are the exact file names you are using and what are the aliases you have defined, it would help us
Good luck
Andreas
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Andreas,
OMP-production on hostname IPCBY004
OMI- Integration on hostname IPCBY104
The alias name is OMDDB.
Now the production is up on ipcby004 with the profiles START_DVEBMGS00_ipcby004
OMP_DVEBMGS00_ipcby004
but we want the system to take the profiles
START_DVEBMGS00_ompdb
OMP_DVEBMGS00_ompdb
We tried changing the .dbenv and .sapenv to ompdb and we tried by replaing the the START_DVEBMGS00_ipcby004 with START_DVEBMGS00_ompdb and changed all the entries to point to ompdb in the start profile instead of ipcby004.
When we do all this and start SAP, we get the message DB startup failed,unable to connect to SQLPLUS.
Regards
Bharathwaj V
Hello Bharathwaj ,
All the changes you have done look ok to me, but they refer to the SAP system. To check what happens with the database, login as orasid in the OMDDB hardware and see if the environment variables are correctly set there. Also make sure that files
.dbenv_omddb.csh and .sapenv_omddb.csh exist under orasid and the hostname they contain is indeed the omddb
If you still get error messages, maybe you could send me the errors in the Oracle alert log? Just do sqlplus "/ as sysdba" from orasid, startup and then send me what is in the alert log
Good luck
Andreas
Hi Andreas,
Because we failed once with the cluster failover thing we tried to bring up OMP on the same host(IPCBY004) with the profiles OMPDB but here also we had the same problem.
The startup of the DB failed and <b>we started DB manually through SQLPLUS and SAP was up but again it took only the profiles of IPCBY004 and not of OMPDB</b>
As we tried it out on the same system we dont have any changes in the ENV and we also made sure that the .dbenv_ompdb.csh and .sapenv_ompdb.csh and all existed.
As the system was up we tried a checkdb and the following is the log
Job started
Step 001 started (program RSDBAJOB, variant &0000000000047, user ID QX64145)
Execute logical command BRCONNECT On host ompdb
Parameters: -u / -c -f check
BR0801I BRCONNECT 7.00 (24)
BR0278E Command output of 'sqlplus':
sh: sqlplus: not found.
BR0280I BRCONNECT time stamp: 2007-09-09 01.16.28
BR0279E Return code from 'sqlplus': 127
BR0302E SQLPLUS call for database instance OMP failed
BR0303E Determination of Oracle version failed
BR0280I BRCONNECT time stamp: 2007-09-09 01.16.28
BR0301E SQL error -12545 at location db_connect-2, SQL statement:
'CONNECT /'
ORA-12545: Connect failed because target host or object does not exist
BR0310E Connect to database instance OMP failed
BR0806I End of BRCONNECT processing: cdwcoekh.log2007-09-09 01.16.28
BR0280I BRCONNECT time stamp: 2007-09-09 01.16.28
BR0804I BRCONNECT terminated with errors
External program terminated with exit code 3
BRCONNECT returned error status E
Job finished
As its our production system we need to try out our other options only this coming weekend ater a successful offline backup.
Regards
Bharathwaj V
Hello,
The "real" error here is that
BR0278E Command output of 'sqlplus':
sh: sqlplus: not found.
Could you go manually in your orasid session and see if you can run sqlplus? Is the $ORACLE_HOME/bin directory present in your system? Is the $ORACLE_HOME/bin directory included in the $PATH variable ?? You can see this by using the env command, check this please in both csh and sh environments
I hope this helps
Andreas
Hi Andreas,
As i had told u once we tried with the cluster failover,we did not try that again and we tried to change the profiles to OMPDB in the same host.
The sqlplus was available at $ORACLE_HOME/bin and the ENV also had it.
There is no change in ENV as we tried it in the <b>same hostname </b> and the .csh and .sh profiles were also the same as
.sapenv_ipcby004.csh were copied to .sapenv_ompdb.csh
.sapenv_ipcby004.sh were copied to .sapenv_ompdb.sh
dbenv_ipcby004.csh were copied to dbenv_ompdb.csh
dbenv_ipcby004.sh were copied to dbenv_ompdb.sh
These were done for both oraomp and ompadm.
Regards
Bharathwaj V
Hi Andreas,
1)Under Normal circumstances,
where we use .dbenv,.sapenv of IPCBY004 and the profiles of IPCBY004,the system comes up fine without any problem in startsap itself.
2)When we change the .dbenv and .sapenv to ompdb and the profiles also to ompdb,that time the startsap failes to dtart the DB saying sqlplus is not available as its not able to find the sqlplus.exe.
But When we switch to oraomp and login into sqlplus,we are able to login and we are able start the DB.
Here we have neither changed the ENV of the ompadm,oraomp nor the path.
Regards
Bharathwaj V
Hi Andreas,
I am using the standard startsap available at /sapmnt/OMP/exe.
There it calls the startsap always with a $hostname in the whole of the script.
ex:PROFILES=`echo /usr/sap/$SAPSYSTEMNAME/SYS/profile/START_SCS*_"$HOSTNAME"`
At OS level we always have the output of hostname as ipcby004
I am planning to give the following lines in a new startsap script and try it this weekend
START_PROFILE="START_DVEBMGS00_ompdb"
DB_SID=OMP
and change all the entried in the START_DVEBMGS00_ompdb to point to ompdb
like
Execute_00 = local $(DIR_EXECUTABLE)/sapmscsa -n pf=$(DIR_PROFILE)/OMP_DVEBMGS00_ompdb
What do u suggest on this.
Do i need to change anything else also????
By the way thanx a lot for being so patient and replying to all questions
Regards
Bharathwaj V
Hi Andreas,
I have changed the Startsap Script this weekend and i have made the system to take the correct profile ompdb.
But i have a small doubt.I have neither changed the dbenv nor the sapenv to ompdb all are on ipcby004.
Now when i do a cluster failover onto the ipcby104 system will the system come up with the correct profile?
Regards
Bharathwaj V
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.