Prev: The Oracle ODBC driver story
Next: Help: How can i audit this
From: Mladen Gogala on 15 Jun 2010 20:27
On Tue, 15 Jun 2010 21:35:17 +0200, Frank van Bortel wrote:


> ASM to ASM cluster

Well, you don't have much choice when we're talking about RAC. I heard
horror stories about OCFS and corruption although I have experienced no
such thing. On the other hand, the company that I work for is maintaining
3 node RAC on ASM for a few years now and we didn't have any problems
with it.
1 select instance_number,instance_name,version,status
2* from gv$instance
SQL> /

INSTANCE_NUMBER INSTANCE_NAME VERSION STATUS
--------------- ---------------- ----------------- ------------
1 ADPRD1 10.2.0.4.0 OPEN
3 ADPRD3 10.2.0.4.0 OPEN
2 ADPRD2 10.2.0.4.0 OPEN


SQL> select group_number,name,state,total_mb
2 from v$asm_diskgroup
3 /

GROUP_NUMBER NAME STATE TOTAL_MB
------------ ------------------------------ ----------- ----------
1 DGDATA CONNECTED 2465760
2 DGFRA CONNECTED 698357
3 DGREDO CONNECTED 147454

SQL> select * from v$version;

BANNER
----------------------------------------------------------------
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bi
PL/SQL Release 10.2.0.4.0 - Production
CORE 10.2.0.4.0 Production
TNS for Linux: Version 10.2.0.4.0 - Production
NLSRTL Version 10.2.0.4.0 - Production

SQL> select inst_id,startup_time from gv$instance;

INST_ID STARTUP_T
---------- ---------
1 28-MAR-10
3 28-MAR-10
2 28-MAR-10

ASM on a RAC cluster functions reliably and fast. I am looking forward to
upgrading to 11g which supports larger allocation units.

--
http://mgogala.byethost5.com
From: Avijit on 22 Jun 2010 20:43
On Jun 11, 12:46 pm, Mladen Gogala <n...(a)email.here.invalid> wrote:
> On Fri, 11 Jun 2010 11:23:50 -0700, charles wrote:
> > Our CRS is at 10.2.0.4, so we should not need toupgradeour
> > clusterware, am i right?  And CRS 11.2 is nowhere to be found anyway.
>
> > Your suggestion is highly appreciated.
>
> CRS 11.2 is on the same OTN location as is the database itself. I
> upgraded the CRS and it went well, without a glitch. The only annoying
> thing is that now you have to ask the network administrator for yet
> another IP address, in addition to the virtual IP's.
>
> --http://mgogala.byethost5.com

Hi Mladen:
I am rellay interested to know how you performed the upgrade? Did you
do a rolling upgrade?
Recently I was trying to test out a rolling upgrade from a 2 node
10.2.04 RAC to 11.2 on RHEL5. I have created a 1 node physical standby
10.2.0.4 and converted it to logical standby. Now my plan is to stop
the logical apply and upgrade it to 11.2 and then start redo apply and
once it is in sync qwith the primary do a switchover.
If you also used the same process can you please share the steps here
so that I can try to move forward.
Thanks.
From: Mladen Gogala on 23 Jun 2010 02:25
On Tue, 22 Jun 2010 17:43:25 -0700, Avijit wrote:

> Hi Mladen:
> I am rellay interested to know how you performed the upgrade? Did you do
> a rolling upgrade?

No. That is not possible with 10.2. There was a hefty downtime involved.

> Recently I was trying to test out a rolling upgrade from a 2 node
> 10.2.04 RAC to 11.2 on RHEL5. I have created a 1 node physical standby
> 10.2.0.4 and converted it to logical standby. Now my plan is to stop the
> logical apply and upgrade it to 11.2 and then start redo apply and once
> it is in sync qwith the primary do a switchover. If you also used the
> same process can you please share the steps here so that I can try to
> move forward.
> Thanks.

Nope. Nothing like that. I shut down both instances, upgraded CRS, then
ASM then database. DBUI gives you a lot of options. The RAC was down
during the whole procedure, which took approximately 3 hours of downtime.
The whole thing went through smoothly.





--
http://mgogala.byethost5.com
From: Avijit on 23 Jun 2010 03:49
On Jun 22, 11:25 pm, Mladen Gogala <gogala.mla...(a)gmail.com> wrote:
> On Tue, 22 Jun 2010 17:43:25 -0700, Avijit wrote:
> > Hi Mladen:
> > I am rellay interested to know how you performed theupgrade? Did you do
> > a rollingupgrade?
>
> No. That is not possible with 10.2. There was a hefty downtime involved.
>
> > Recently I was trying to test out a rollingupgradefrom a 2 node
> > 10.2.04 RAC to 11.2 on RHEL5. I have created a 1 node physical standby
> > 10.2.0.4 and converted it to logical standby. Now my plan is to stop the
> > logical apply andupgradeit to 11.2 and then start redo apply and once
> > it is in sync qwith the primary do a switchover. If you also used the
> > same process can you please share the steps here so that I can try to
> > move forward.
> > Thanks.
>
> Nope. Nothing like that. I shut down both instances, upgraded CRS, then
> ASM then database. DBUI gives you a lot of options. The RAC was down
> during the whole procedure, which took approximately 3 hours of downtime.
> The whole thing went through smoothly.
>
> --http://mgogala.byethost5.com

THanks for your reply.
Just a few clarifications:
1. did u shutdown the 10g ASM and db instances and the crs before u
started the upgrade? Asking because, when I tried to run the
runInstaller from 11g grid installation location it complained during
configuration of SCAN that the vip is already assigned. But when I
stopped the 10g CRS it did not complian and proceeded
2. will running runInstaller from grid installation directory upgrade
both COS and ASM?...while starting runInstaller, it automatically
selects the option "upgrdae grid infrastructure", does it mean both
CRS and ASM?
Thanks
From: Mladen Gogala on 23 Jun 2010 10:04
On Wed, 23 Jun 2010 00:49:56 -0700, Avijit wrote:

> 1. did u shutdown the 10g ASM and db instances and the crs before u
> started the upgrade?

DB & ASM were shut down, CRS was running.

> Asking because, when I tried to run the
> runInstaller from 11g grid installation location it complained during
> configuration of SCAN that the vip is already assigned. But when I
> stopped the 10g CRS it did not complian and proceeded 2. will running
> runInstaller from grid installation directory upgrade both COS and
> ASM?...while starting runInstaller, it automatically selects the option
> "upgrdae grid infrastructure", does it mean both CRS and ASM?
> Thanks

Yes.



--
http://mgogala.byethost5.com
First  |  Prev  |  Next  |  Last
Pages: 1 2 3
Prev: The Oracle ODBC driver story
Next: Help: How can i audit this