Mosa3lyan

updated

15 years ago

M

Mosa3lyan

2 Intern

•

265 Posts

0

1488

March 25th, 2012 00:00

OpenReplicator with Oracle error

after i migrate using OR hot push migration  i face issue with oracle,
start copy from storage system while  oracle online , for all disks with consistent mode ( which freez IO then start  copying for all disk for oracle rac on aix gpfs cluster at the same time)
after copying reach 100%
i asked them to stop oracle and umount all disks
and i make incremental ( recreate) copy with consistent to make sure source and  destiniation disks are 100% in sync
they have 2 database on that cluster , one of them go up and the other one  have problem with below error
do u know the cause , or how to fix it
thanks in advanc
eC:\Documents and  Settings\eccadmin\Desktop\VMAX>symcli

Symmetrix Command Line Interface  (SYMCLI) Version V7.3.1.0 (Edit Level: 1305)
built with SYMAPI Version  V7.3.1.0 (Edit Level: 1305)we are migrate data between DMX and vmax using open replicator for aix hosts  whick has 2 database.

one database is up and working fine but second  database we are getting error
below error messages which we are receiving  while opening the OMSTST database .

Database mounted in Shared Mode  (CLUSTER_DATABASE=TRUE)
Lost write protection disabled
Completed: ALTER  DATABASE MOUNT /* db agent *//* {1:32065:341} */
ALTER DATABASE OPEN /* db  agent *//* {1:32065:341} */
This instance was first to open
Abort recovery  for domain 0
Errors in file  /oracle/apps/base/diag/rdbms/omstst/OMSTST1/trace/OMSTST1_ora_1171466.trc:
ORA-01122:  database file 1 failed verification check
ORA-01110: data file 1:  '/oradata1/OMSTST/system01.dbf'
ORA-01207: file is more recent than control  file - old control file
ORA-1122 signalled during: ALTER DATABASE OPEN /* db  agent *//* {1:32065:341} */...
Mon Mar 19 16:21:18 2012
Shutting down  instance (abort)
License high water mark = 2
USER (ospid: 606430):  terminating the instance
Instance terminated by USER, pid = 606430
Mon Mar  19 16:21:19 2012
Instance shutdown complete
  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    997

    0

    Posted March 25th, 2012 20:00

    OR steps are fine, it's either Oracle or cluster config related.

  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    739

    0

    Posted March 25th, 2012 06:00

    can you post your OR commands ?

  • Mosa3lyan

    2 Intern

    •

    265 Posts

    741

    0

    Posted March 25th, 2012 07:00

    yep

    1-Oracle online

    create and activate

    2-shoutdown oracle and umount

    recrerate , activate then termintae

    should i shutdown oracle in first activate or only for incremental or copy

  • Mosa3lyan

    2 Intern

    •

    265 Posts

    739

    0

    Posted March 25th, 2012 07:00

    should i put oracle in hot backup mode , although i dont think it will make a difference as  i use -consistent

  • Mosa3lyan

    2 Intern

    •

    265 Posts

    739

    0

    Posted March 25th, 2012 07:00

    symrcopy -f elian.txt create -copy -hot -push -differential

    symrcopy -f elian.txt activate -consistent

    symrcopy -f elian.txt query

    symrcopy –file elian.txt set pace 2 

    symrcopy list ceiling -dir all -sid 2221   #  all are none so pace value will be effective

    symrcopy -f elian.txt query -details

    symrcopy -f elian.txt verify

    once copy reach 100% ,

    Stop the application/DB/Cluster  , GPFS AIX CLUSTER with oracle rac and umount all san disks

    symrcopy -file elian.txt recreate

    # symrcopy -file elian.txt activate -consistent

    then i make sure invalide tracks=0 and reach 100%

    symrcopy -f dev_file_OR query -details

    symrcopy -f dev_file_OR verify

    # symrcopy -file dev_file_OR terminate

    thats it

    then bingbong error on oracle control file

  • Mosa3lyan

    2 Intern

    •

    265 Posts

    739

    0

    Posted March 25th, 2012 07:00

    ya it should , its really strange to have control files not sync with data file

    i

  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    739

    1

    Posted March 25th, 2012 07:00

    this should work , the only thing that i do differently:

    I let it seat in this precopy mode until the migration

    symrcopy -f create -push -hot -differential -name oraclemigration1 -precopy -pace 5

    shutdown oracle host

    symrcopy -f activate -consistent

  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    741

    0

    Posted March 25th, 2012 07:00

    but you already shutdown the database before you do the activate right ?

  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    741

    0

    Posted March 25th, 2012 08:00

    i assume you shutdoown the  host, activate OR, remove old devices, present new devices ?

  • dynamox

    11 Legend

    •

    20419 Posts

    •

    87439 Points

    741

    0

    Posted March 25th, 2012 08:00

    Mosa3lyan wrote

    should i shutdown oracle in first activate or only for incremental or copy

    should only need to shut it down for final cut-over.