Re: ASM Question/Suggestions - moving from 10TB devices to 20TB devices and diskgroup move

Good point. I was actually not aware of the WAIT clause. Good to know.

On Wed, Sep 8, 2021 at 11:55 AM Mikhail Velikikh <mvelikikh_at_gmail.com> wrote:

> NOTE: Make *sure* the rebalance is *complete* before attempting to remove
>> the old LUNs! In other words, be sure that V$ASM_OPERATION is empty before
>> proceeding to remove the old device from the server.
>>
>
> I would add that we could use the WAIT clause for that:
>
> https://docs.oracle.com/en/database/oracle/oracle-database/19/sqlrf/ALTER-DISKGROUP.html#GUID-22D73AB6-7063-4627-A2ED-18D521ED2557
>
>> Use this clause to specify when, in the course of the rebalance
>> operation, control should be returned to the user.
>>
>> -
>>
>> Specify WAIT if you want control returned to the user after the
>> rebalance operation has finished. You can explicitly terminate a rebalance
>> operation running in WAIT mode, although doing so does not undo any
>> completed disk add, drop, or resize operations in the same statement.
>>
>> e.g. (ADD/DROP/REBALANCE in one command):
>
>> ALTER DISKGROUP "<name>"
>> ADD DISK '<path>' NAME "<name>"
>> DISK '<path>' NAME "<name>"
>> DROP DISK "<old_disk>"
>> DISK "<old_disk>" *REBALANCE POWER 11 WAIT*
>>
>> Or we can use it in a separate rebalance operation too:
> https://docs.oracle.com/en/database/oracle/oracle-database/19/sqlrf/ALTER-DISKGROUP.html#GUID-22D73AB6-7063-4627-A2ED-18D521ED2557
>
> [image: image.png]
>
>
> On Wed, 8 Sept 2021 at 16:09, Mark J. Bobak <mark_at_bobak.net> wrote:
>
>> I agree with Mikhail. That is how I've done it in the past, and it
>> worked well.
>>
>> NOTE: Make *sure* the rebalance is *complete* before attempting to
>> remove the old LUNs! In other words, be sure that V$ASM_OPERATION is empty
>> before proceeding to remove the old device from the server.
>>
>> -Mark
>>
>> On Wed, Sep 8, 2021 at 10:51 AM Mikhail Velikikh <mvelikikh_at_gmail.com>
>> wrote:
>>
>>> Sorry, missed the list in my reply.
>>>
>>> Yes, doing it in one command is the best approach. It rebalances the
>>> data only once.
>>>
>>>
>>> https://docs.oracle.com/en/database/oracle/oracle-database/19/ostmg/alter-diskgroups.html#GUID-6BB31112-8687-4C1E-AF14-D94FFCDA736F
>>>
>>>> You can add, resize, or drop disks while the database remains online.
>>>> Whenever possible, multiple operations in a single ALTER DISKGROUP
>>>> statement are recommended. Grouping operations in a single ALTER
>>>> DISKGROUP statement can reduce rebalancing operations.
>>>>
>>> There will be no rebalance at all if you resize the existing disks
>>> (option 1 in my previous email). It is possible if you use the same storage.
>>> Of course, you have to drop the extra disks that you don't need which
>>> roughly rebalances just a half of data.
>>>
>>>
>>>
>>> On Wed, 8 Sept 2021 at 15:28, Chris Taylor <
>>> christopherdtaylor1994_at_gmail.com> wrote:
>>>
>>>> Ok, so I think I'm following and that makes sense.
>>>>
>>>> In a nutshell it would be add all the new 20T devices *AND* drop the
>>>> existing 10T devices in the same command so that the rebalance only
>>>> rebalances to the 20T devices?
>>>>
>>>> Is that the theory/idea?
>>>>
>>>> That would definitely prevent the need to use rman to move the
>>>> datafiles or anything.
>>>>
>>>> Chris
>>>>
>>>>
>>>> On Wed, Sep 8, 2021 at 10:09 AM Mikhail Velikikh <mvelikikh_at_gmail.com>
>>>> wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I know ASM devices should be the same size when you rebalance.
>>>>>>
>>>>>
>>>>> If it is an EXTERNAL redundancy disk group, then you can have a disk
>>>>> of different size there.
>>>>> 12c ASM: Unable To Add New Disks With Dissimilar Size To 12.1.0.2 ASM
>>>>> Diskgroups (Normal or High Redundancy) Due To ORA-15410 (New 12c ASM
>>>>> Enhancement Validation/Constraint) (Doc ID 1938950.1)
>>>>> <https://support.oracle.com/rs?type=doc&id=1938950.1>
>>>>> Although it is not a best practice, it can be used during such
>>>>> migrations.
>>>>>
>>>>> Is that the *only* option here? Anyone have additional
>>>>>> suggestions/thoughts?
>>>>>>
>>>>>
>>>>> 1. You can resize the existing disks online, then drop the disks you
>>>>> don't need.
>>>>>
>>>>> https://docs.oracle.com/en/database/oracle/oracle-database/19/ostmg/alter-diskgroups.html#GUID-6AEFFA72-7BDC-4AA8-8667-8417AAF3DAC8
>>>>> It didn't work for me two years ago with AFD. I haven't encountered
>>>>> the same issue with ASMlib or without any Oracle drivers.
>>>>> If you don't use AFD, ignore notes below:
>>>>> Cannot resize ASM Disk using ASM Filter Driver(AFD) online (Doc ID
>>>>> 2099164.1)
>>>>> ASM not displaying resized disk size in AFD configuration (Doc ID
>>>>> 2263094.1)
>>>>> Bug 29893839 - ASM CRASHES WHEN RUNNING AFD_REFRESH AFTER THE LUN
>>>>> RESIZE AT OS LAYER (Doc ID 29893839.8)
>>>>>
>>>>> 2. You can provision new 20TB LUNs and DROP/ADD disk in one ALTER
>>>>> DISKGROUP command. ASM automatically rebalances data.
>>>>>
>>>>> https://docs.oracle.com/en/database/oracle/oracle-database/19/ostmg/alter-diskgroups.html#GUID-14E48E96-F5D6-4F41-A8E4-B535CBD3CAE5
>>>>>
>>>>> Best regards,
>>>>> Mikhail Velikikh
>>>>>
>>>>>
>>>>>
>>>>> On Wed, 8 Sept 2021 at 14:46, Chris Taylor <
>>>>> christopherdtaylor1994_at_gmail.com> wrote:
>>>>>
>>>>>> List,
>>>>>>
>>>>>> We have a need to restructure our underlying ASM devices to 20T from
>>>>>> 10T to reduce the number of luns assigned to a host.
>>>>>>
>>>>>> The Total DiskGroup SIze is 163840.00 GB (~ 160 TB) made up of 10TB
>>>>>> devices. (16 Devices)
>>>>>>
>>>>>> Our storage admins would like to move these to 20T devices and halve
>>>>>> the number of physical devices presented.
>>>>>>
>>>>>> I know ASM devices should be the same size when you rebalance.
>>>>>>
>>>>>> I know *one* option is to add the 8 20T devices as candidates and
>>>>>> create a new diskgroup and move the data to the new diskgroup and then drop
>>>>>> the old 10T devices from ASM.
>>>>>>
>>>>>> Is that the *only* option here? Anyone have additional
>>>>>> suggestions/thoughts?
>>>>>>
>>>>>> Chris
>>>>>>
>>>>>>

--
http://www.freelists.org/webpage/oracle-l





image/png attachment: image.png