Re: ASM Dependency on VIP and 10.2.0.3
- From: Naqi Mirza <naqimirza@xxxxxxxxx>
- To: Alex Gorbachev <gorbyx@xxxxxxxxx>
- Date: Mon, 9 Apr 2007 14:51:36 -0700 (PDT)
Thanks for the reply. Unfortunately I do not have an environment to test this
out on yet (thats another story). With that said, I noticed something from the
crs_stat -p command taken from an old 10g (10.2.0.2) installation. For the
database instances, their resources have a parameter called:
.
REQUIRED_RESOURCES=ora.node_nameN.vip ora.node_nameN.ASM1.asm
.
Even if the workaround for asm to remove its dependency on the vip is done,
seems that the rdbms instance will still go offline. The whole point of
applying the workaround is to avoid having to restart the asm and rdbms
instance (of course the workaround is only for asm to begin with). Anyway, I
initially thought the db instance went down because its asm instance went
offline - seems that is not the case.
.
You mention the use of setting check_interval to 0. If this is done, and
checking of the vip resource is disabled, I assume that means it will not
relocate the vip either? - which would mean back to long delays (pre 10g RAC)
of network timeouts for clients no?
----- Original Message ----
From: Alex Gorbachev <gorbyx@xxxxxxxxx>
To: naqimirza@xxxxxxxxx
Cc: oracle-l@xxxxxxxxxxxxx
Sent: Tuesday, 10 April, 2007 2:17:25 AM
Subject: Re: ASM Dependency on VIP and 10.2.0.3
I was hitting this problem as far bac as in 10.1.0.3 and at that time
spent countless hours for Oracle support and development trying to
figure out the best solution. Removing dependency wasn't recommended
by support even though development was quite sure it should work. And
it did work when we tested it.
The best workaround recommended was to set check interval to 0 for VIP
resources that ultimately disables checking completely. Thus, public
network failures are not affecting VIP resource status but any manual
manipulation of VIP resources (stopping) would be obviously reflected
on ASM and databases.
I quickly tried it in 10.2.0.3 and it seems to be working OK but do
you own homework.
I posted about it some time ago if interested -
http://www.pythian.com/blogs/384/vip-dependencies-lifted-in-oracle-10203
On 4/9/07, Naqi Mirza <naqimirza@xxxxxxxxx> wrote:
>
> Hi,
> Currently running 10.2.0.2 RAC with ASM. I have 2 questions (hoping someone
> has experienced this):
>
> 1. A public network failure results in a relocation of the vip. The
> relocation of the vip results in the corresponding asm instance shutting
> down, which in turn stops the rdbms instance associated with it. There is a
> workaround in metalink, that discusses how to remove this dependency (for
> releases under 10.2.0.3).
> I was wondering if anyone has used this workaround? I also wanted to know if
> once this dependency has been removed, and there is a public network failure
> - once this (public network failure) is resolved does the vip automatically
> relocate back to its original node or does this still need to be done
> manually. Apparently in 10.2.0.3 once the public network issue is fixed the
> vip will automatically relocate back - since asm no longer has a dependency
> it and the rdbms instance will not need to be started.
>
> 2. Can or Has anyone tested public network failures in 10.2.0.3 - does it
> work as advertised ?
>
> Thanks
>
> ________________________________
> New Yahoo! Mail is the ultimate force in competitive emailing. Find out
> more at the Yahoo! Mail Championships. Plus: play games and win prizes.
--
Best regards,
Alex Gorbachev
http://www.oracloid.com
___________________________________________________________
Copy addresses and emails from any email account to Yahoo! Mail - quick, easy
and free. http://uk.docs.yahoo.com/trueswitch2.html
Other related posts:
