Oracle 10g Release 2 RAC with ASM on HP-UX 11.11

[Wachyu Danyanto]

Hi,

I have a problem during creation of ASM diskgroup in Cluster Environment (Oracle 10g Release 2) on HP-UX 11.11.

I have some questions related to Oracle 10g R2 RAC with ASM on HP-UX 11.11:
1. If ASM going to used for storage of Oracle Database, does it need ServiceGuard to be activated? If it doesn't need ServiceGuard, why I cannot create ASM Diskgroup pointing to the raw disk (/dev/rdsk/c6t0d0) ?

The errors say:
ORA-15018: diskgroup cannot be created
ORA-15030: diskgroup name "DGSYSTEM" is in use by another diskgroup

2. I have another error when try to running srvctl:

$ srvctl config
PRKH-1010 : Unable to communicate with CRS services.
[PRKH-1000 : Unable to load the SRVM HAS shared library
[PRKN-1008 : Unable to load the shared library "srvmhas10"
or a dependent library, from
SHLIB_PATH="/apps/oracle/product/10.2/crs/jdk/jre/lib/PA_RISC2.0:/apps/oracle/product/10.2/crs/jdk/jre/lib/PA_RISC2.0/server:/apps/oracle/product/10.2/crs/jdk/jre/../lib/PA_RISC2.0:/apps/oracle/product/10.2/crs/lib32:/apps/oracle/product/10.2/crs/srvm/lib32:/apps/oracle/product/10.2/db/lib32:/apps/oracle/product/10.2/db/rdbms/lib32:/usr/lib"
[java.lang.UnsatisfiedLinkError: /apps/oracle/product/10.2/crs/lib32/libsrvmhas10.sl: specified library, or one of its dependencies, does not exist.]]]


Could anyone help me with this?
Does anyone have the same experiences issues like this?

Thanks in advance.

Regards,
Danyanto

[Wachyu Danyanto]

Hello,

Is this forum active?

Regards,
Danyanto

[Zinky]

Yes we're alive.

The non-response can only mean one thing -- possibly no one's using ASM in this forum.

I myself have doubts using ASM unless the storage, sysadmin and dba groups are part of a homogenous organization.

Say -- what kind of array are you using with your ASM implementation?

[Wachyu Danyanto]

"
Yes we're alive.

The non-response can only mean one thing -- possibly no one's using ASM in this forum.

I myself have doubts using ASM unless the storage, sysadmin and dba groups are part of a homogenous organization.

Say -- what kind of array are you using with your ASM implementation? "

Hi,

I believe in HP/Oracle CTC, there are someone have tried Oracle 10g R2 RAC with ASM on HP-UX 11.23.
Because they have write cookbook about installation of Oracle 10g R2 on HP-UX 11.23.

About the storage, we are using HP Storage 12000 (OPENV).

About the operating system, we used HP-UX 11.11.

Is there any problem with ASM without ServiceGuard in HP-UX 11.11?

Thank you.

Regards,
Danyanto

[Zinky]

You will be better off running RAC on HP-UX 11i v2 (11.23). Veritas components on HP-UX 11.11 are pretty dated (based on VxVM 3.5/VxFS 3.5).

Also, you have the perfect array - XP12000 for your cluster/RAC/ASM implementation. ASM considers Hitachi based storage frames like the XP12000 as one of its fully certified storage arrays.

If you are testing/playing Oracle RAC with ASM implementation -- PLEASE read through the requirements/manuals very thoroughly.


Selamat Tembang.

[Zinky]

Please see also:

http://www2.itrc.hp.com/service/james/dispDoc.do?docURL=http%3A%2F%2Fsearch.hp.com%2Fredirect.html%3Furl%3Dhttp%253A%2F%2Fdocs.hp.com%2Fen%2F8831%2F10gRAC%26qt%3D%252BXP12000%26hit%3D19&aid=SEARCH_MANUAL&pil=17&serStr=XP12000

[Duncan Edmonstone]

Generally you don't use ASM with Serviceguard - if you want to use ASM you use a complete Oracle stack - no HP clustering products at all *unless* you want to use SLVM - in that case you can use ASM with Serviceguard.

As to the other messages - I'm no RAC DBA, but I would certainly fix your CRS problem before attempting *anything* else. IIRC ASM relieas on CRS so you need to fix that problem before touching ASM. A quick google on the problem seems to indicate this can arise when libclntsh shared lib isn't created in CRS_HOME. Do you have a libclntsh.sl file in CRS_HOME/network/lib ?

The ASM problem may just be a symptom on the CRS problem, or it could be that ASM is finding existing ASM headres on the disk in question. If thats the case and you're *completely sure* the disk isn't in use by any system, then just writing over the start of the disk should let ASM use it:

dd if=/dev/zero of=/dev/rdsk/c6t0d0 bs=8192 count=1000

HTH

Duncan

[Wachyu Danyanto]

Hi Nelson,

Thank you for your reply.
About the technical reference Guide about Oracle 10g RAC using SGeRAC. But unforunately we are not using ServiceGuard for the ASM. Only raw device. ServiceGuard packages is already installed but not configured yet.

About the operating system, Oracle said it compatible with the HP-UX 11i v1.
Is it possible not to upgrade the Operating System to 11.23?

Thank you.

Regards,
Danyanto

[Duncan Edmonstone]

So you are using Oracle RAC 10g with Oracle's own clusterware, but you are *also* going to use Serviceguard on this system as well?

I think you need to get this checked with both youre Oracle and HP support people, I doubt either would be happy with two seperate cluster managers running on the same nodes - who is responsible for fencing?


HTH

Duncan

[Wachyu Danyanto]

"
Generally you don't use ASM with Serviceguard - if you want to use ASM you use a complete Oracle stack - no HP clustering products at all *unless* you want to use SLVM - in that case you can use ASM with Serviceguard.

As to the other messages - I'm no RAC DBA, but I would certainly fix your CRS problem before attempting *anything* else. IIRC ASM relieas on CRS so you need to fix that problem before touching ASM. A quick google on the problem seems to indicate this can arise when libclntsh shared lib isn't created in CRS_HOME. Do you have a libclntsh.sl file in CRS_HOME/network/lib ?

The ASM problem may just be a symptom on the CRS problem, or it could be that ASM is finding existing ASM headres on the disk in question. If thats the case and you're *completely sure* the disk isn't in use by any system, then just writing over the start of the disk should let ASM use it:

dd if=/dev/zero of=/dev/rdsk/c6t0d0 bs=8192 count=1000

HTH

Duncan "

Hi Duncan,

Yes, we are planned that we don't use ServiceGuard for the ASM, only raw disk devices.

About the libclntsh.sl , yes it is there.
CRS is running well, but after I patched the clusterware to 10.2.0.3, it seems that it have a problem with the srvctl utility.

Before I created diskgroup in the ASM, I have dd all the raw disk devices. All of it from /dev/rdsk/c6t0d0 - /dev/rdsk/c6t0d3.
But when creating Diskgroup it showing error
like this:

create diskgroup dgdata02 external redundancy disk '/dev/rdsk/c6t0d2'
*
ERROR at line 1:
ORA-15018: diskgroup cannot be created
ORA-15030: diskgroup name "DGDATA02" is in use by another diskgroup

I have tried with all of the raw disk, it all show the same error.

This ASM problem occur first time before I patch to 10.2.0.3. I read from Oracle Metalink article about this ASM error and I was think this is bug from Oracle 10.2.0.1, so I patched the Oracle clusterware and database software to 10.2.0.3. But after been patched, the Error on ASM still don't solve yet, still showing the same error and also another error is come when I want to use srvctl utility.

I have repeatedly installed Oracle clusterware and Database many many times, but the ASM Problem didn't solve yet.
The Production server needs to go alive soon after ASM is created then I need to configuring RAC Database on ASM and then the production database server will be ready for used.

Really need help on this.
Thank you very much.

Regards,
Danyanto