fouaddba
asked on
Upgrade to 10gR2 from 9.2
Hi guys,
I am doing an upgrade from 9.2 to 10gr2, box is solaris 10.
Have already installed 10gr2 and its working fine there, a small db have already created there using exp/imp,
but since yesterday I have started an upgrade 60 gb of db, everyting works fine,
have startup upgrade command,
created the new tbsp AUX and then executed the upgrade script:
@catupgrade.sql
yesterday 05/17 at 3:00 pm and since now it is still running, can't say what is the problem,
is there anyone who is aware of this kind of problem, that this script hanged for so long...
giving an excertp from the alert.log file: tail -f alert.log
tail of alert.log
Thu May 18 07:28:45 2006
Incremental checkpoint up to RBA [0x4ec.11a25.0], current log tail at RBA [0x4ec.11a25.0]
Thu May 18 07:59:05 2006
Incremental checkpoint up to RBA [0x4ec.11a44.0], current log tail at RBA [0x4ec.11a44.0]
Thu May 18 08:29:25 2006
Incremental checkpoint up to RBA [0x4ec.11a65.0], current log tail at RBA [0x4ec.11a65.0]
Thu May 18 08:59:45 2006
Incremental checkpoint up to RBA [0x4ec.11a82.0], current log tail at RBA [0x4ec.11a87.0]
Thu May 18 09:30:04 2006
Incremental checkpoint up to RBA [0x4ec.11aa7.0], current log tail at RBA [0x4ec.11aa7.0]
a little from begining of alert.log
ALTER SYSTEM SET resource_manager_plan='' SCOPE=MEMORY;
Wed May 17 15:47:27 2006
replication_dependency_tra cking turned off (no async multimaster replication found)
Completed: ALTER DATABASE OPEN MIGRATE
Wed May 17 15:49:54 2006
CREATE TABLESPACE sysaux DATAFILE '/db/db01/oradata/stg1/sys aux01.dbf'
SIZE 500M REUSE
EXTENT MANAGEMENT LOCAL
SEGMENT SPACE MANAGEMENT AUTO
ONLINE
Wed May 17 15:50:05 2006
Completed: CREATE TABLESPACE sysaux DATAFILE '/db/db01/oradata/stg1/sys aux01.dbf'
SIZE 500M REUSE
EXTENT MANAGEMENT LOCAL
SEGMENT SPACE MANAGEMENT AUTO
ONLINE
Wed May 17 16:11:40 2006
Beginning log switch checkpoint up to RBA [0x4e9.2.10], SCN: 4317919789
Thread 1 advanced to log sequence 1257
Current log# 5 seq# 1257 mem# 0: /db/db01/oradata/stg1/redo 05a.log
Current log# 5 seq# 1257 mem# 1: /db/db01/oradata/stg1/redo 05b.log
Wed May 17 16:16:51 2006
Completed checkpoint up to RBA [0x4e9.2.10], SCN: 4317919789
Wed May 17 16:17:52 2006
Require any one assistance, who has experienced of this kind of behaviour,
thanks in advance...
I am doing an upgrade from 9.2 to 10gr2, box is solaris 10.
Have already installed 10gr2 and its working fine there, a small db have already created there using exp/imp,
but since yesterday I have started an upgrade 60 gb of db, everyting works fine,
have startup upgrade command,
created the new tbsp AUX and then executed the upgrade script:
@catupgrade.sql
yesterday 05/17 at 3:00 pm and since now it is still running, can't say what is the problem,
is there anyone who is aware of this kind of problem, that this script hanged for so long...
giving an excertp from the alert.log file: tail -f alert.log
tail of alert.log
Thu May 18 07:28:45 2006
Incremental checkpoint up to RBA [0x4ec.11a25.0], current log tail at RBA [0x4ec.11a25.0]
Thu May 18 07:59:05 2006
Incremental checkpoint up to RBA [0x4ec.11a44.0], current log tail at RBA [0x4ec.11a44.0]
Thu May 18 08:29:25 2006
Incremental checkpoint up to RBA [0x4ec.11a65.0], current log tail at RBA [0x4ec.11a65.0]
Thu May 18 08:59:45 2006
Incremental checkpoint up to RBA [0x4ec.11a82.0], current log tail at RBA [0x4ec.11a87.0]
Thu May 18 09:30:04 2006
Incremental checkpoint up to RBA [0x4ec.11aa7.0], current log tail at RBA [0x4ec.11aa7.0]
a little from begining of alert.log
ALTER SYSTEM SET resource_manager_plan='' SCOPE=MEMORY;
Wed May 17 15:47:27 2006
replication_dependency_tra
Completed: ALTER DATABASE OPEN MIGRATE
Wed May 17 15:49:54 2006
CREATE TABLESPACE sysaux DATAFILE '/db/db01/oradata/stg1/sys
SIZE 500M REUSE
EXTENT MANAGEMENT LOCAL
SEGMENT SPACE MANAGEMENT AUTO
ONLINE
Wed May 17 15:50:05 2006
Completed: CREATE TABLESPACE sysaux DATAFILE '/db/db01/oradata/stg1/sys
SIZE 500M REUSE
EXTENT MANAGEMENT LOCAL
SEGMENT SPACE MANAGEMENT AUTO
ONLINE
Wed May 17 16:11:40 2006
Beginning log switch checkpoint up to RBA [0x4e9.2.10], SCN: 4317919789
Thread 1 advanced to log sequence 1257
Current log# 5 seq# 1257 mem# 0: /db/db01/oradata/stg1/redo
Current log# 5 seq# 1257 mem# 1: /db/db01/oradata/stg1/redo
Wed May 17 16:16:51 2006
Completed checkpoint up to RBA [0x4e9.2.10], SCN: 4317919789
Wed May 17 16:17:52 2006
Require any one assistance, who has experienced of this kind of behaviour,
thanks in advance...
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
i hope it is so,
but I can not see SHARED_POOL and JAVA_POOL size.
In 10g there is a famous fat Listener bug:
This is the typical 10g trap.
Go to the remote 10g installation.
In ...\network\admin open listener.ora
see
SID_LIST_LISTENER =
(SID_LIST =
(SID_DESC =
(SID_NAME = PLSExtProc)
(ORACLE_HOME = D:\Ora10g)
(PROGRAM = extproc)
)
(SID_DESC =
(SID_NAME = test)
(ORACLE_HOME = D:\Ora10g)
)
)
Add the entry
(SID_DESC =
(SID_NAME = test)
(ORACLE_HOME = D:\Ora10g)
)
with regard to the particular Oracle Home and SID Name.
Restart the listener:
c:>lsnrctl stop
c:>lsnctl start
Thats all.
After fixing this make sure that you connections to the instance run smootly.
Restart the instance and try again. check again the prerequisites fopr running catupgrade.sql
.
Create a log file and see the progress:
SQL>spool /u03/upgr.log
SQL>@ .../catupgr.sql
SQL>spool off
If there are errors you will see them in the log also.
I just upgraded 3 10g R2 Instances on Linux - had no problems at all!
but I can not see SHARED_POOL and JAVA_POOL size.
In 10g there is a famous fat Listener bug:
This is the typical 10g trap.
Go to the remote 10g installation.
In ...\network\admin open listener.ora
see
SID_LIST_LISTENER =
(SID_LIST =
(SID_DESC =
(SID_NAME = PLSExtProc)
(ORACLE_HOME = D:\Ora10g)
(PROGRAM = extproc)
)
(SID_DESC =
(SID_NAME = test)
(ORACLE_HOME = D:\Ora10g)
)
)
Add the entry
(SID_DESC =
(SID_NAME = test)
(ORACLE_HOME = D:\Ora10g)
)
with regard to the particular Oracle Home and SID Name.
Restart the listener:
c:>lsnrctl stop
c:>lsnctl start
Thats all.
After fixing this make sure that you connections to the instance run smootly.
Restart the instance and try again. check again the prerequisites fopr running catupgrade.sql
.
Create a log file and see the progress:
SQL>spool /u03/upgr.log
SQL>@ .../catupgr.sql
SQL>spool off
If there are errors you will see them in the log also.
I just upgraded 3 10g R2 Instances on Linux - had no problems at all!
Check also Kernel and Root user prerequisites for Solaris Oracle installations.
ASKER
thanks schwertner,
but this is the status:
. first I tired with manul upgrade, it stucked at this location and doesn't move at all...
-- Load all the Java classes
begin if initjvmaux.startstep('CREA TE_JAVA_SY STEM') then
2 initjvmaux.rollbacksetup;
this is the location where it stucked when I first asked the question, next day I killed the session and
tried to do it with dbua, was looking a the log and the same thing happen,
after coming to 73%, it stucked at the same location:
-- Load all the Java classes
begin if initjvmaux.startstep('CREA TE_JAVA_SY STEM') then
2 initjvmaux.rollbacksetup;
we left yesterday, and today still it was there, stoped it and starts again,
now unchecking the OEM repository box and again it went fine,
but tell you one thing it was keep giving some sys.... table constraint violation errors...and I was ignoring it,
then it reached at 85% and I can see it is the same place, is it amazing....as usuall nothing GOOD response from Oracle support.
-- Load all the Java classes
begin if initjvmaux.startstep('CREA TE_JAVA_SY STEM') then
2 initjvmaux.rollbacksetup;
Can't say what to do...?
but this is the status:
. first I tired with manul upgrade, it stucked at this location and doesn't move at all...
-- Load all the Java classes
begin if initjvmaux.startstep('CREA
2 initjvmaux.rollbacksetup;
this is the location where it stucked when I first asked the question, next day I killed the session and
tried to do it with dbua, was looking a the log and the same thing happen,
after coming to 73%, it stucked at the same location:
-- Load all the Java classes
begin if initjvmaux.startstep('CREA
2 initjvmaux.rollbacksetup;
we left yesterday, and today still it was there, stoped it and starts again,
now unchecking the OEM repository box and again it went fine,
but tell you one thing it was keep giving some sys.... table constraint violation errors...and I was ignoring it,
then it reached at 85% and I can see it is the same place, is it amazing....as usuall nothing GOOD response from Oracle support.
-- Load all the Java classes
begin if initjvmaux.startstep('CREA
2 initjvmaux.rollbacksetup;
Can't say what to do...?
ASKER
and one more to add, the alert log, in all the 3 instances looked liked this:
SERVER COMPONENT id=CATPROC: timestamp=2006-05-19 09:55:35
Fri May 19 09:56:02 2006
Incremental checkpoint up to RBA [0x4f2.13999.0], current log tail at RBA [0x4f2.1a114.0]
Fri May 19 09:56:40 2006
SERVER COMPONENT id=RDBMS: status=VALID, version=10.2.0.1.0, timestamp=2006-05-19 09:56:40
Fri May 19 09:57:32 2006
create rollback segment MONSTER2 storage (initial 2 m next 2 m maxextents unlimited)
Fri May 19 09:57:32 2006
Completed: create rollback segment MONSTER2 storage (initial 2 m next 2 m maxextents unlimited)
Fri May 19 09:57:32 2006
alter rollback segment MONSTER2 online
Completed: alter rollback segment MONSTER2 online
what do you think the problem is...
SERVER COMPONENT id=CATPROC: timestamp=2006-05-19 09:55:35
Fri May 19 09:56:02 2006
Incremental checkpoint up to RBA [0x4f2.13999.0], current log tail at RBA [0x4f2.1a114.0]
Fri May 19 09:56:40 2006
SERVER COMPONENT id=RDBMS: status=VALID, version=10.2.0.1.0, timestamp=2006-05-19 09:56:40
Fri May 19 09:57:32 2006
create rollback segment MONSTER2 storage (initial 2 m next 2 m maxextents unlimited)
Fri May 19 09:57:32 2006
Completed: create rollback segment MONSTER2 storage (initial 2 m next 2 m maxextents unlimited)
Fri May 19 09:57:32 2006
alter rollback segment MONSTER2 online
Completed: alter rollback segment MONSTER2 online
what do you think the problem is...
What do you mean saying MANUAL UPGRADE?
Of the DB?
The best way to do Upgrade is Export/Import way!
I have posted many instructions here how to do this.
Of the DB?
The best way to do Upgrade is Export/Import way!
I have posted many instructions here how to do this.
ASKER
no I know, that is the best way and I did it too for my small db, 45gb,
but these 2 are 1.2tb & 2.9tb, they are pretty big,
that's the reason was using first manual upgarde and then dbua...but none worked...:-(
but these 2 are 1.2tb & 2.9tb, they are pretty big,
that's the reason was using first manual upgarde and then dbua...but none worked...:-(
If you have Metalink account call Oracle Support.
Also plan export/import way - schema by schema.
It is the most reliable way.
Also plan export/import way - schema by schema.
It is the most reliable way.
ASKER
thanks, I am also in contact with them:
SR Number 5436801.992 Open Date 18-MAY-06 14:35:03
Support Identifier 1547029 Name Syed Hashim
Severity 4 Last Update 22-MAY-06 14:32:17
Product Oracle Server - Enterprise Edition Product Version 10.2.0.1.0
Platform Solaris Operating System (SPARC 64-bit) Detailed Status 1st Callback
SR Reference n/a BUG Reference n/a
but as usual nothing happend yet....
SR Number 5436801.992 Open Date 18-MAY-06 14:35:03
Support Identifier 1547029 Name Syed Hashim
Severity 4 Last Update 22-MAY-06 14:32:17
Product Oracle Server - Enterprise Edition Product Version 10.2.0.1.0
Platform Solaris Operating System (SPARC 64-bit) Detailed Status 1st Callback
SR Reference n/a BUG Reference n/a
but as usual nothing happend yet....
I have indications that the upgrade is very long process.
May be you have to start it and to wait patiently for result.
Nevertheles try to backup the DB for security reason.
May be you have to start it and to wait patiently for result.
Nevertheles try to backup the DB for security reason.
ASKER
but more then 17 hours for a small db of size 60gb... I don't think so,
and the most important all the 3 sessions hung/stuck at the same place for that much long,
still struggling with Oracle support, but I think got to resolve it by myself, lets see;
-- Load all the Java classes
begin if initjvmaux.startstep('CREA TE_JAVA_SY STEM') then
2 initjvmaux.rollbacksetup;
and same type of messeges in the alert.log too, there must be something wrong,
where, can't say at this point in time...
what do you suggest over here, what is the best way to resolve this situation, because I know that I can do exp/imp,
but now I don't want to, want to upgrad w/out using exp/imp.
and thanks for your concern, schwertner, you always upgrade using exp/imp or ...?
and the most important all the 3 sessions hung/stuck at the same place for that much long,
still struggling with Oracle support, but I think got to resolve it by myself, lets see;
-- Load all the Java classes
begin if initjvmaux.startstep('CREA
2 initjvmaux.rollbacksetup;
and same type of messeges in the alert.log too, there must be something wrong,
where, can't say at this point in time...
what do you suggest over here, what is the best way to resolve this situation, because I know that I can do exp/imp,
but now I don't want to, want to upgrad w/out using exp/imp.
and thanks for your concern, schwertner, you always upgrade using exp/imp or ...?
May be you have an upgraded Oracle version (9i, 10g) but you have upgraded only the software. After upgrading the software you have also to upgrade the Dictionary.
If this is not done usually the DBA got a defect instance.
Seems that your java machine is not working.
The prerequisites
JAVA_POOL at least 150m
same for SHARED_POOL is very important.
If this is not done usually the DBA got a defect instance.
Seems that your java machine is not working.
The prerequisites
JAVA_POOL at least 150m
same for SHARED_POOL is very important.
ASKER
SQL> show sga
Total System Global Area 1593835520 bytes
Fixed Size 1978976 bytes
Variable Size 771755424 bytes
Database Buffers 805306368 bytes
Redo Buffers 14794752 bytes
SQL>
so you are suggesting, to shutdown this process and starts again...