Comments on: Voting Disk Redundancy in ASM

By: jee

jee — Sun, 19 Aug 2012 05:46:44 +0000

Very good explanation. In context to exadata, if no of db nodes are 2 & cell nodes are more than 3….What will be the scenariot of DB nodes are 2 & cell nodes are only 3….and CD with high redudancy…I mean VD will be allocated & survive on disk failure.

By: Andy Colvin

Andy Colvin — Mon, 30 Jan 2012 20:28:00 +0000

Santhosh,

I haven’t attempted this, but I believe that if you lost 2 failgroups simultaneously, your diskgroup would go offline. The reason that I expect that is because you have the option to store other files in the diskgroup, so ASM would want to protect the files that could possibly be stored there. I’ll test this and let you know what I find.

By: Santhosh

Santhosh — Mon, 30 Jan 2012 06:40:57 +0000

Hi ,

Great post. Have one clarification though … I am in process of designing a RAC with two separate storage devices.

we are trying to eliminate single storage device failure leading to RAC or database failure.

If we have five failure groups in the disk group(normal redundancy) holding voting disk and two of the disk holding voting disk goes offline due to storage failure.

Will ASM recreate the voting disks or will this scenario lead to cluster / node re-boots.

Thanks you.

By: Andy Colvin

Andy Colvin — Mon, 09 Jan 2012 15:38:53 +0000

disk_repair_time should be visible in v$asm_attribute. I have a script for that in my scripts section (asm_attributes.sql).

By: Fairlie Rego

Fairlie Rego — Mon, 09 Jan 2012 06:46:28 +0000

Sorry got it from the md_backup command

@diskgroup_set = (
{
‘ATTRINFO’ => {
‘_._DIRVERSION’ => ’11.2.0.2.0′,
‘DISK_REPAIR_TIME’ => ’0.5H’,

By: Fairlie Rego

Fairlie Rego — Fri, 06 Jan 2012 23:43:58 +0000

Gr8 post.. As an aside after changing the disk_repair_time I can’t seem to locate the new value in any of the ASM views. Hopefully I am missing something simple.