Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

1106

August 4th, 2010 13:00

Creating a new pool & start replicating

Hi,

Currently we are facing an issue, one of the FAB application using the default pool, along with other applications. Now they are asking us to setup a replication for FAB.

option 1: Can we move the FAB data from default pool to a new pool, by doing archieving ?

option 2: Need to setup entire default pool replication ?

Requesting you to  please provide the commands to implement above options.

Thanks in advance !!

Surya,

Storage Engineer

27 Posts

August 4th, 2010 16:00

Hello Surya,

I notice that there are two profiles ' admin ' and ' anonymous ' that writes to the default pool.  You will have to check in the application as of what profile the FAB application uses. In case all the applications use the same anonymous profile, I have the steps that you need to follow to ensure that application writes to it's own pool.

Start Pool Migration

This procedure describes how to migrate data from the default pool to a custom pool. The pool migration is a two-step process. First, the scratchpad mappings are copied to the active mappings and then the migration process starts on each node asynchronously.

When to Use this Procedure

If you have an application which writes data to the default pool and you want to change it to write to its own pool, you first must migrate existing data to the new pool. Data written to the default pool will remain there when the application starts using its own pool unless the data is migrated.

Before You Start

The following steps must be completed before data can be migrated:

  1. Create the custom pool for the application with the create pool command.


  2. Create or update the profile that the application will use to write to the new pool with the create profile or update profile command.


  3. Map the application profile to the new pool with the create poolmapping command.


  4. Optionally check the pool mappings with the show pool mapping command.


  5. Optionally check the pool capacity with the show pool capacity command.


  6. Migrate the data from the default pool to the new pool as described below.


When using replication, ensure that you issue the same active mapping on all clusters involved in the replication topology.

Migrating Data from the Default Pool

Start the CLI.

Connect as a user with the accesscontrol role.

Issue the command:

migrate poolmapping start

  1. Type y to confirm that you wish to start the migration based on the scratchpad mappings:


Do you want to start a migration process based on the scratchpad mappings? (yes, no) [no]: y

You are prompted whether to change the information about pool capacity to reflect any changes in pool sizes.

Capacity information might become incorrect

Do you want to recalculate it? (yes, no) [yes]: y

Type y and press Enter to accept the option.

Type y and press Enter to issue the command and start the migration:

Issue the Command? (yes, no) [no]: y

Committed the scratchpad mappings and started the migration process.

27 Posts

August 4th, 2010 15:00

My previous post should have started of like

'Before we proceed you should understand that if you have used a single profile for all the applications to write to the Centera or if you do not have profiles created for the corresponding applications I do not see any possibility to replicate data written only by the FAB application. You will have to replicate the entire pool.'

27 Posts

August 4th, 2010 15:00

Before we proceed you should understand that if you have used a single profile for all the applications to write to the Centera or if do you have profiles created for the corresponding applications I do not see any possibility to replicate data written only by the FAB application. You will have to replicate the entire pool. If not you can set replication on the profile used by the FAB application.

C-Clips can be moved based on the profile that originally wrote the C-Clip.  If different applications were writing to the default pool using different profiles, the C-Clips can be moved to application specific pools.  If they both used the same profile, including the anonymous profile, they can be moved but not separated.


The purpose of this procedure is to explain how to migrate C-Clips from the default pool to a custom application pool. C-Clips can be migrated provided they were written using the SDK 2.3 or higher.

CAUTION: EMC discourages to use pool migration as a general practice. Contact your EMC representative if you want move data from the default pool to a custom application pool.

Note : C-Clips written using version 2.2 SP2 or earlier will always belong to the default pool. It is not possible to migrate these C-Clips to an application pool.

Note : This procedure consumes system resources and you may notice an impact on performance whilst it is running.  The migration task is, however, gated to 10 C-Clips/second per node so the impact will be minimal.

Note : When a migration task is completed, the active mapping must remain on the cluster - this allows C-Clips to retain their new pool membership when regenerated or restored.

Procedure

Note : Follow this procedure on each cluster in the replication topology to ensure that the pool ownership on each cluster is consistent.

1. Launch the CLI and connect to the source cluster.

2. Add a mapping between the profile used to write the C-Clips to the default pool and a pool using the create poolmapping   command.  The pool needs to exist on the cluster for the command to work and cannot be the default pool.

3. The mappings are placed on the scratchpad which is used to prepare mappings until the user has the correct mappings to start the actual migration.

4. View the scratchpad pool mapping to check that it is correct using the show pool mapping command.

5. Follow the procedure from Step 2 for each profile with legacy data in the default pool.

6. Once the mappings are in place, activate the pool mapping and start the migration task using the migrate poolmapping start command. This command copies all scratchpad mappings to the active mappings and re-starts the migration process.

7. Monitor the migration task using the show pool migration command.

38 Posts

August 4th, 2010 15:00

Thanks Radha.

Please find the below profiles listed on Centera.

Config# show profile list
Profile Name              Home Pool       Enabled  Configured Usage
-----------------------------------------------------------------------------
anonymous                 default         yes      Data Access, Management
admin                     default         yes      Management
afi-ivr                   afi-ivr         yes      Data Access, Management
CFA1pro                   CFA1            yes      Data Access
CFA2pro                   CFA2            yes      Data Access, Management
console                   ConsoleArchive  yes      Data Access, Management
REP0243                   cluster         yes      Data Access, Management
temp                      CFA2            yes      Data Access, Management
-----------------------------------------------------------------------------

Not sure whether each application using a seperate profile.  If so then we can move the data from default pool  to seperate pool, as you mentioned later we can initiate the replication.

38 Posts

August 9th, 2010 13:00

Thanks alot Radha !!

We have Implemented the change, created the new pool, profile provided the PEA file to customer & established the replication as well.

No Events found!

Top