Does using ASM meaningful to LUN of SAN?

When use local disk or disk array device, we add disk to ASM, ASM strip and balance the disks.
But when using LUN of SAN, LUN already strip or mirrored by SAN, what is ad for using ASM?
Thanks.

The advantage is the same. ASM stripes every datafile over every LUN in the ASM diskgroup, whether that LUN is composed of local disk, direct attached disk, or disk on a SAN.
There are numerous whitepapers on this topic out there.
http://www.oracle.com/technology/products/database/asm/index.html
Regards,
Greg Rahn
http://structureddata.org

Similar Messages

Should I partition the disks using fdisk (300 GB lun) for ASM?

We've configured oracle 11g and asm on SLES10sp2 systems prior to the one I'm working on now. We use disks that are luns from SAN storage. The SAN devices vary (Tier 1 and tier 2). In earlier installs, we did not need to partition the disks using fdisk (luns up to 40GB in size). We did need to write a label using fdisk and 'w'. The asm that is being configured now is using 300 GB luns (on sles10 sp2 oracle 11g). I've needed to create a new primary partition using fdisk and then have asm 'look' as the partitioned device. If I did not do that, ASM would not look at the unpartitioned device.
The database in now built and it seems to be working ok.
Does this config sound ok?
Thanks,
Eleanor

I think this is right. I also always had to partition disks and let asm use the partitions because for some reason it did not accept the full device.
Bjoern

New 11.2 database instalation using ASM

Hello peeps,
My unix team has asked me how many mount points i would like on the unix system? Normally i use OFA of /u01, /u02, /u03, /u04
With this new database i will be installing dataguard using ASM. So my question is : what do i tell the Unix team for mount points?
Does ASM just use one mount point /u01?

Hi,
The answer might not have helped you as such. ASM is totally different from what you were used to in normal NFS or RAW. It talkes a lot more to configure and you need to know what you are doing. If this is a production system, I would suggest that you get hold of a decent book on ASM as the guide suggested might not be adequate for a full fledge Prod installation for an ASM newbie. ASM allows you to present LUNs or disks to the database is a much more different manner than you might be used to.
Regards,
T

Advice on whether to use ASM for 11gR2 databases

Hi all,
Hope I can get some advice here for ASM.
I'm going to upgrade all my 10.2 EE databases to 11gR2, and I will not be running RAC, only single instances and Data Guard. My storage is HP SAN. I wonder if it is worth the trouble (in our small shop) to run ASM.
Currently, my backups are done from a SAN perspective only, meaning that whatever is in the SAN storage is backed up.
How do I back up the files in ASM? Do I have to copy the backupsets to file system first ?
Thanks!
June

fjfranken wrote:
For small databases performance gain is zero, you only add complexity where a normal file system would be sufficient.Disagree.
Also, the reasoning is not valid. ASM works at device level. Not at cooked file system level.
This means that ASM directly uses the LUN (seen as a local scsi device) of the SAN. See LUN. Use LUN. That simple.
Not using ASM means that the o/s needs to be updated and configured. That LUN needs to be mounted. It needs to be formatted. The mount needs to be made permanent (via fstab or whatever).
This makes ASM less complex and cooked file system more complex.
Also consider that the mount and its contents are exposed to other s/w and processes and users. This raises issue such as access control and security - and the possibility of errant s/w trashing a database file because it happens to be running on the wrong mount or wrong directory.
This is not the case with an ASM diskgroup as it is not mounted and not visible as a normal cooked file system.
And if a cooked file system feature needed for a non-database files, an ACFS volume can be created in the diskgroup and mounted as a clustered file system. So it is not as if you loose the option of having a cooked file system on a LUN, when using ASM.
Database size has nothing to do with the decision between cooked file system or ASM when using SAN LUNs. Ease of use, manageability, and feature set, have. From this perspective a cooked file system for database storage is old and clunky and insecure, and ASM is flexible, performant, scalable and secure.

Can't create database using ASM (SOLVED)

Hi all
I'm trying to use ASM for the first time, on Oracle 10.2.0.1 on Solaris x64.
I have installed the ASM instance into /opt/oracle/asm/10.2.0 and created disk groups. I have cssd running OK. I am able to start and stop the ASM instance without problems, and I can select from v$asm_diskgroup to confirm that disks are mounted OK.
I have then installed Oracle EE separately into /opt/oracle/server/10.2.0. I first did a software only install, and now I am trying to create a DB.
The problems come when I try to use this ASM instance to host a new database. I first tried to use DBCA to create a new database, but on database creation I got the following errors:
ORA-00200: control file could not be created
ORA-00202: control file: '+DBLIVE1'
ORA-17502: ksfdcre:4 Failed to create file +DBLIVE1
ORA-15001: diskgroup "DBLIVE1" does not exist or is not mounted
ORA-15055: Message 15055 not found; No message file for product=RDBMS, facility=ORA
ORA-01031: insufficient privileges
I then told DBCA just to create the DB creation scripts, and I tried manually running these with SQL*Plus.
When doing it with SQL*PLus, I initially got the same error as shown above. But then something changed (sorry, not sure what), and now the error I get is:
CREATE DATABASE "NEONREL1"
ERROR at line 1:
ORA-01501: CREATE DATABASE failed
ORA-00349: failure obtaining block size for '+DBLIVE1'
ORA-01031: insufficient privileges
I've put some debug info below, showing me succesfully connecting to the ASM instance and then attempting to create the DB using the db creation scripts, showing the error at the end. You can see that the oracle OS user is able to connect fine to ASM, then I swithc ORACLE_SID and ORACLE_HOME to the EE install and try to create the DB, at which point it apparently can't connect to ASM any more.
I've tried the DB creation many times, and in between attempts I completely empty $ORACLE_HOME/admin/<dbname> and delete the files related to the attempted install from $ORACLE_HOME/dbs/ . I've also stopping/starting ASM, rebooting, and I've done the install of ASM and EE a couple of times over in case I made any mistakes in my earlier attempts.
Any help would be much appreciated!
Tom
##### CHECKING ASM
oracle@neonrcom-db1:~$ uname -a
SunOS neonrcom-db1 5.10 Generic_127128-11 i86pc i386 i86pc
# css is running
oracle@neonrcom-db1:~$ ps -ef | grep css
oracle 498 1 0 21:46:40 ? 0:01 /opt/oracle/asm/10.2.0/bin/ocssd.bin
# listener is running in the ASM instance
oracle@neonrcom-db1:~$ ps -ef | grep tnsl
oracle 1332 1 0 21:49:59 ? 0:00 /opt/oracle/asm/10.2.0/bin/tnslsnr LISTENER -inherit
# ASM is only entry in /var/opt/oracle/oratab
oracle@neonrcom-db1:~$ grep -v "^#" /var/opt/oracle/oratab
+ASM:/opt/oracle/asm/10.2.0:N
# I can connect to ASM fine, and it has diskgroups mounted.
oracle@neonrcom-db1:~$ export ORACLE_HOME=/opt/oracle/asm/10.2.0
oracle@neonrcom-db1:~$ export ORACLE_SID='+ASM'
oracle@neonrcom-db1:~$ sqlplus "sys as sysdba"
SQL*Plus: Release 10.2.0.1.0 - Production on Mon Jul 21 20:53:10 2008
Copyright (c) 1982, 2005, Oracle. All rights reserved.
Enter password:
Connected to:
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production
With the Partitioning, OLAP and Data Mining options
SQL> set line 150
SQL> select name, block_size, state, type, total_mb, free_mb from v$asm_diskgroup;
NAME BLOCK_SIZE STATE TYPE TOTAL_MB FREE_MB
DBARCH1 4096 MOUNTED EXTERN 2096856 2096784
DBLIVE1 4096 MOUNTED EXTERN 4193904 4193812
#### Contents of init.ora for new DB
db_create_file_dest=+DBLIVE1
db_recovery_file_dest=+DBARCH1
db_recovery_file_dest_size=2147483648
##### DB INSTALLATION ATTEMPT
oracle@neonrcom-db1:~$ export ORACLE_HOME=/opt/oracle/server/10.2.0
oracle@neonrcom-db1:~$ export ORACLE_SID='NEONREL1'
oracle@neonrcom-db1:~$ export PATH=$ORACLE_HOME/bin:$PATH
oracle@neonrcom-db1:~$ /opt/oracle/server/10.2.0/admin/NEONREL1/scripts/NEONREL1.sh
You should Add this entry in the /var/opt/oracle/oratab: NEONREL1:/opt/oracle/server/10.2.0:Y
SQL*Plus: Release 10.2.0.1.0 - Production on Mon Jul 21 22:10:54 2008
Copyright (c) 1982, 2005, Oracle. All rights reserved.
specify a password for sys as parameter 1
Enter value for 1: xxx
specify a password for system as parameter 2
Enter value for 2: xxx
specify a password for sysman as parameter 3
Enter value for 3: xxx
specify a password for dbsnmp as parameter 4
Enter value for 4: xxx
specify ASM SYS user password as parameter 6
Enter value for 6: xxx
Connected to an idle instance.
SQL> spool /opt/oracle/server/10.2.0/admin/NEONREL1/scripts/CreateDB.log
SQL> startup nomount pfile="/opt/oracle/server/10.2.0/admin/NEONREL1/scripts/init.ora";
ORACLE instance started.
Total System Global Area 1.9294E+10 bytes
Fixed Size 2054976 bytes
Variable Size 2264925376 bytes
Database Buffers 1.7012E+10 bytes
Redo Buffers 14721024 bytes
SQL> CREATE DATABASE "NEONREL1"
2 MAXINSTANCES 8
3 MAXLOGHISTORY 1
4 MAXLOGFILES 16
5 MAXLOGMEMBERS 3
6 MAXDATAFILES 100
7 DATAFILE SIZE 300M AUTOEXTEND ON NEXT 10240K MAXSIZE UNLIMITED
8 EXTENT MANAGEMENT LOCAL
9 SYSAUX DATAFILE SIZE 120M AUTOEXTEND ON NEXT 10240K MAXSIZE UNLIMITED
10 SMALLFILE DEFAULT TEMPORARY TABLESPACE TEMP TEMPFILE SIZE 20M AUTOEXTEND ON NEXT 640K MAXSIZE UNLIMITED
11 SMALLFILE UNDO TABLESPACE "UNDOTBS1" DATAFILE SIZE 200M AUTOEXTEND ON NEXT 5120K MAXSIZE UNLIMITED
12 CHARACTER SET AL32UTF8
13 NATIONAL CHARACTER SET UTF8
14 LOGFILE GROUP 1 SIZE 51200K,
15 GROUP 2 SIZE 51200K,
16 GROUP 3 SIZE 51200K
17 USER SYS IDENTIFIED BY "&&sysPassword" USER SYSTEM IDENTIFIED BY "&&systemPassword";
CREATE DATABASE "NEONREL1"
ERROR at line 1:
ORA-01501: CREATE DATABASE failed
ORA-00349: failure obtaining block size for '+DBLIVE1'
ORA-01031: insufficient privileges
Message was edited by:
tjobbins

Update: I've worked out the difference between the two sets of errors I get.
The basic error is this:
ORA-00200: control file could not be created
ORA-00202: control file: '+DBLIVE1'
ORA-17502: ksfdcre:4 Failed to create file +DBLIVE1
ORA-15001: diskgroup "DBLIVE1" does not exist or is not mounted
ORA-15055: Message 15055 not found; No message file for product=RDBMS, facility=ORA
ORA-01031: insufficient privileges
However if my init.ora contains the line:
control_files=/opt/oracle/server/10.2.0/dbs/cntrlNEONREL1.dbf
then I instead get the second error:
CREATE DATABASE "NEONREL1"
ERROR at line 1:
ORA-01501: CREATE DATABASE failed
ORA-00349: failure obtaining block size for '+DBLIVE1'
ORA-01031: insufficient privileges
So basically these must be the same error, just in the second case I'm not trying to put the control file on the ASM so it fails at a different point.
But both errors must be because of the same cause, I suppose.

Can Grid Control 10gR2 create a standby database when primary db uses ASM?

Does anyone know if Grid Control 10gR2 will be able to create a standby database when the primary database uses ASM for data files, redo log files, and archive log files but the standby database will use mounted disks? If it can, would it matter that the primary db will be a RAC cluster and the standby will be a single instance db?

I appreciate your reply and hope you are correct. Have you had a look at the 10gR2 Grid Control to confirm this or are you only going on published statements? The reason I ask is because previous versions of Grid Control claimed to support ASM but they would only allow you to work with them AFTER you had created a standby database manually but you couldn't actually use Grid Control to create a standby database from a primary database that had its log files on ASM disks.

Oracle11g R2 Active Data guard using ASM Problem?

I have configured oracle11g r2 RAC on 2 notes using ASM Grid ( OS unix).
RAC is up and running.
Now I am configuring Active data Guard.
Under grid user instance +ASM and listener is running.
Under oracle user static listener is running.
All disk is mounted.
Oracle RAC and Data Guard directory and structure I have keeped same.
Now my problem is below:
$ ./rman target sys/HPinvent123nbl@dcpdb AUXILIARY sys/HPinvent123nbl@drpdb
Recovery Manager: Release 11.2.0.1.0 - Production on Wed Jan 16 16:28:32 2013
Copyright (c) 1982, 2009, Oracle and/or its affiliates. All rights reserved.
connected to target database: DCPDB (DBID=316773134)
connected to auxiliary database: DRPDB (not mounted)
RMAN> duplicate target database for standby from active database;
Starting Duplicate Db at 16-JAN-13
using target database control file instead of recovery catalog
allocated channel: ORA_AUX_DISK_1
channel ORA_AUX_DISK_1: SID=5644 device type=DISK
contents of Memory Script:
backup as copy reuse
targetfile '/u02/app/oracle/product/11.2.0/dbhome_1/dbs/orapwdcpdb1' auxiliary format
'/u02/app/oracle/product/11.2.0/dbhome_1/dbs/orapwdrpdb' ;
executing Memory Script
Starting backup at 16-JAN-13
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=1897 instance=dcpdb1 device type=DISK
Finished backup at 16-JAN-13
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of Duplicate Db command at 01/16/2013 16:28:48
RMAN-06136: ORACLE error from auxiliary database: ORA-00200: control file could not be created
ORA-00202: control file: '+data'
ORA-17502: ksfdcre:4 Failed to create file +data
ORA-15001: diskgroup "DATA" does not exist or is not mounted
ORA-15055: unable to connect to ASM instance
ORA-01031: insufficient privileges
RMAN>
Please help.\
Thanks
Solaiman

root@drpdb1 []# id oracle
uid=108(oracle) gid=700(oinstall) groups=701(dba)
root@drpdb1 []# id grid
uid=109(grid) gid=700(oinstall) groups=701(dba),702(asmdba)
Edited by: 876149 on Jan 16, 2013 3:19 AM

Using ASM in a cluster, and new snapshot feature

Hi,
I'm currently studying ASM and trying to find some answers.
From what I understand and experienced so far, ASM can only be used as a local storage solution. In other words it cannot be used by network access. Is this correct?
How is the RDBMS database connecting to the ASM instance? Which process or what type of connection is it using? It's apparently not using listener, although the instance name is part of the database file path. How does this work please?
How does ASM work in a cluster environment? How does each node in a cluster connect to it?
As of 11g release 2, ASM provides a snapshot feature. I assume this can be used for the purpose of backup, but then each database should use it's own diskgroup, and I will still need to use alter database begin backup, correct?
Thanks!

Markus Waldorf wrote:
Hi,
I'm currently studying ASM and trying to find some answers.
From what I understand and experienced so far, ASM can only be used as a local storage solution. In other words it cannot be used by network access. Is this correct?Well, you are missing one point that it would entirely depend on the architecture that you are going to use. If you are going to use ASM for a single node, it would be available right there. If installed for a RAC system, ASM instance would be running on each node of the cluster and would be managing the storage which would be lying on the shared storage. The process ASMB would be responsible to exchange messages, take response and push the information back to the RDBMS instance.
How is the RDBMS database connecting to the ASM instance? Which process or what type of connection is it using? It's apparently not using listener, although the instance name is part of the database file path. How does this work please?Listener is not need Markus as its required to create the server processes which is NOT the job of the ASM instance. ASM instance would be connecting the client database with itself immediately when the first request would come from that database to do any operation over the disk group. As I mentioned above, ASMB would be working afterwards to carry forward the request/response tasks.
How does ASM work in a cluster environment? How does each node in a cluster connect to it? Each node would be having its own ASM instance running locally to it. In the case of RAC, ASM sid would be +ASMn where n would be 1, 2,... , going forward to the number of nodes being a part of teh cluster.
As of 11g release 2, ASM provides a snapshot feature. I assume this can be used for the purpose of backup, but then each database should use it's own diskgroup, and I will still need to use alter database begin backup, correct?You are probably talking about the ACFS Snapshot feature of 11.2 ASM. This is not to take the backup of the disk group but its like a o/s level backup of the mount point which is created over the ASM's ACFS mount point. Oracle has given this feature so that you can take the backup of your oracle home running on the ACFS mount point and in the case of any o/s level failure like someone has deleted certain folder from that mountpoint, you can get it back through the ACFS snapshot taken. For the disk group, the only backup available is metadata backup(and restore) but that also does NOT get the data of the database back for you. For the database level backup, you would stillneed to use RMAN only.
HTH
Aman....

Is it better to use ASM if my DB uses only one disk ?

Hello
I am planning to install a NEW database 11g R1 on a new windows server.
I am running a 9.2 DB and will export its data to the freshly installed 11gR1.
I ma wondering if i have to use the ASM feature ?
The dabatase files are small : 52 Gb, and it will keep on one disk.
I have understood that the main benefit of using ASM in automatic multi disk storage.
Does it worth to use it on my small DB ?
Thank you for your kind answers.
Christian

If you want to learn ASM then using it with a single disk system and just creating the diskgroups to not use Oracle redundancy is an option. Also one of the things ASM is designed to provide is async IO, that is, easy managment of raw partitions. Depending on your platform ASM may be desirable from this point of view.
Also does a one disk set-up today mean only a one disk set-up tomorrow? It is possible to migrate a database to a new disk server via ASM while the database is running. You create a mirror, allow it to populate, and then remove the original disk. The database data files are now on a new disk storage unit.
But ASM is by no means required or necessary. You just want to consider all the options it provides and the headaches it brings to determine which set of values: use or not, is right for your situation.
HTH -- Mark D Powell --

Trouble using asm instance for new database

I have an asm instance and am trying to create a new database that will use the asm storage. I'm using dbca to create the new database. This is a test system on windows so I'm using the asmtool -create flatfiles for disks with the asmallow_only_raw_disks parameter set to false.
If I have #SQLNET.AUTHENTICATION_SERVICES = (NTS) commented out in my sqlnet.ora, then I can't get into asmcmd:
ORA-01031: insufficient privileges (DBD ERROR: OCISessionBegin)
and dbca won't create the new database that uses asm because the dbca gets the same 01031 error when trying to create the control files for that database.
If I have SQLNET.AUTHENTICATION_SERVICES = (NTS) NOT commented out, the I can get into asmcmd, but I can't startup the databases using sys, even though I am sitting at the server,
AND, dbca won't create the new database that uses asm becuase the dbca can't start the instance
credential retrieval failure.
I'm in a catch 22. Something is not right with my system but I just can't seem to figure out what it is.
The asm instance does have a password file.

I let the dbca create scripts and am currently walking through the scripts. I think it's my asm database that is actually creating the problem. I did the orapwd and oradim manually and can startup nomount the instance just fine. The create database statement fails on creating the controlfile but it's saying it already exists. I suspected the dbca form wasn't giving me all the errors like it should.
Thanks for your suggestion on the password file.

How to make a physical copy of database when using ASM

I used to make a physical copy of production database
every night to create test environment. I have
simply physically copied files, recreated control file,
... to create it.
But on 10g I want to use ASM, is there some way
to do physical copy of the database similar way?

You can not directly to duplicate a regular database that does not use ASM into new one that does not use ASM. The problem with ASM is that you can not see the files directly at the disk level.
If you use RMAN you could use the command duplicate and all would be easier.
Joel PÃ©rez
http://www.oracle.com/technology/experts

Installing Grid Control and not using ASM

Is it possible to install Grid Control and not be using ASM? We are on the brink of beginning our install (Peoplesoft) and may not use ASM. It seems to me from documentation that I'm reviewing that installing GC assumes you will be using ASM.
Any info would be appreciated.

Hi,
the Grid Control Repository database does not require ASM. It works perfectly fine with datafiles on a normal filesystem.
Best regards,
Robert

Any one used ASM CP command to do an endian conversion

using Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production on Solaris. We want to convert it to Linux. We have read about the asm cp command doing the conversion, This will allow us to bypass using RMAN Convert.
Does anyone have any experience doing this?

Hi,
Are two different things.
ASM is a filesystem with own endian, which is platform dependent.
Datafile is a File which have also its own endian and is stored over ASM Disk (Filesystem) which have its own endian.
Example:
Datafile of my database on AIX have Big Endian stored on ASM Disk (Filesystem) with Big Endian.
But I can convert this Datafile to Little Endian on AIX (i.é RMAN Convert datafile on source before move to destination) and store the Datafile Little Endian over ASM Disk with Big Endian format.
So, in this case I can use ASM CP from Big ENDIAN to move datafile converted (Little endian) to my destination which is ASM with Little Endian.
How to Migrate to different Endian Platform Using Transportable Tablespaces With RMAN (Doc ID 371556.1)
If the source files are on ASM diskgroup and target platform has different endianess , the CONVERT command needs to be executed. The files cannot be copied directly between two ASM instances of different platform.

User-managed backups of a 10g DB using ASM

Is it possible to do user-managed backups of a 10g database which is using ASM?
When I run the following SQL command
SELECT NAME FROM v$datafile;
It returns
+DBDATA/db_test/datafile/system.256.617633575
+DBDATA/db_test/datafile/users.259.617633757
The ASM is setup in 'normal redundency' across 2 partitions, both of which are not formatted.
If there is anyone who has done user-mangaged backups on a 10g DB using ASM who could provide an insight I would be greatful.
Basically what I would like to know is how do I access the datafiles, so I can back them up to a location of my choosing?

I think it's a good product, we use it for many years now. A big advantage is Backtrack automates not only backups (that do all products), it also provides automatic restore/recovery in many cases (not in all!). That means the product detects what is wrong with the database and automatically performs all necessary steps to bring back the database to working. In case of trouble that's good for operators by night and also good for the DBA, who does not need to come.
Disadvantage (may be true for all 3rd party products) Backtrack is always behind Oracle's newest technology, so it took some time to support ASM and it's still not a full support.
There's no problem with online backups, Backtrack uses the traditional way (ALTER TABLESPACE BEGIN BACKUP .... END BACKUP), they don't use RMAN technology in the background.
We use both in our company, Backtrack and RMAN. Besides the technical point of view, there's also a license issue. Backtrack's license fee is due to amount of data, the larger your databases the larger the fee.
Werner

How to migrate RAC/ASM 10g to a new SAN on Unix

Hi everybody,
I have an Oracle RAC 10g using ASM on Solaris Sparc 10.
And I need to change my SAN storage for a better.
Somebody can tell me the steps to follow or a Metalink note?
Thanks & regards...

Hi
With ASM your life is simple.. Just add the new disks from new storage to ASM and drop old disks and throttle the reblance operations..
ALTER DISKGROUP <DG> Add disk <from new storage>
ALTER DISKGROUP <DG) drop disk <old disks>
ALTER DISKGROUP <DG> REBALANCE POWER 11;
You can query the V$ASM_OPERATIONS to monitor the rebalance progress. Once all the disks are dropped, you can unplug the old storage from the system.
Similarly, you can use ocrconfig -replace ocr/ocrmirror to migrate the OCR (Online)
Votedisks, you need to bring the cluster down (due to a bug in 10g, online if you are 11)..
# crsctl query css votedisk
# crsctl add css votedisk <new vote disk> /crsctl delete css votedisk <old vote disk>

Does using ASM meaningful to LUN of SAN?

Similar Messages

Maybe you are looking for