cancel
Showing results for 
Search instead for 
Did you mean: 

How to use Slave Repository?

siowfong_chen
Contributor
0 Kudos

Hi! I am reading about the creation of slave repository and have successfully created one in our MDM system. Now, I found that first of all the master and slave cannot be loaded at the same time. Is that right?

So, I unloaded the master and loaded the slave. Then when I do a shop transaction (SRM-MDM), the system is not able to connect to the slave automatically. I found that I have to change the catalog naming in the SRM configuration to point to this slave repository. Does that mean that should we decide to use slave repository, we should assign this repository to the users and not the master repository?

Appreciate some clarification on this.

Cheers!

SF

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Hi,

As Jitesh mentioned Master Slave concept is used for Staging and Data Availability in case of SRM MDM.

Yes you will have to assign the Slave repository to the user as system automatically does not recognise the difference.

You can Refer note no

1170807 and also1180802 for more clarifiaction.

Regards

Nisha

siowfong_chen
Contributor
0 Kudos

Thanks Nisha, the notes are indeed very helpful. I am able to amend the web service and the system can now pick the correct repository automatically.

Cheers!

SF

Answers (3)

Answers (3)

Former Member
0 Kudos

Hi SF,

If you want to know how to use swap the repositories, you can refer the below link

http://help.sap.com/saphelp_mdm550/helpdata/en/8e/9f9c427055c66ae10000000a155106/frameset.htm

CLIX Command Line Interface -> Command Detail -> Repository Control Commands -> Master/Slave Commands.

There is a concept of Repository PreLoad which loads the repository without assigning any port and then use the Repository Swap command to Swap the Master with the Slave one.

Regards,

Jitesh Talreja

Former Member
0 Kudos

Hi SF,

Master-Slave configuration is mainly used for backup purpose in which Slave is the read only copy of the Master repository. When the Master is down, slave can take over its part (atleast for data reading) unless master repository is up. This type of configuration gives the following advantages (From reference guide):

1. Performance boosts

They reduce load and improve performance by distributing read-access requests among multiple servers.

2. Data availability

They ensure that repository data is available even when the master repository is busy.

3. Redundancy

They reduce the risk of repository downtime due to server outages.

4. Staging

They allow you to make incremental changes on a u201Cstagingu201D repository and then release the changes only when you are ready.

5. Efficient updating

They speed up data synchronization by sending only repository updates, not entire repositories.

As far as SRM-MDM scenario is concerned, i dont think this would be useful since we need to pass the repository name in the webservice. What else can be done is create the webservice for Slave as well but dont assign this catalog to Org Plan. When the Master repository is down, just assign this catalog to Org Plan and remove the Master one.

For more information, refer the below link

http://help.sap.com/saphelp_mdm550/helpdata/en/8e/9f9c427055c66ae10000000a155106/frameset.htm

Repository Administration -> Maintaining Master and Slave Repository

Regards,

Jitesh Talreja

siowfong_chen
Contributor
0 Kudos

Thanks. May be I was confused since I was trying to apply the functionalities to SRM-MDM.

Regards

SF

Former Member
0 Kudos

Hi SF,

One more thing is even in normal operations there would not be automatic switching between Master and Slave. Usually what happens is there is a concept of Port Swapping(using CLIX) which put slave repository on the same port on which Master was running so that users should not face any down time.

Regards,

Jitesh Talreja

Former Member
0 Kudos

Hi SF,

By default the operation always points to the Master repository in a Master slave landscape.

As its the master repository where all the changes and configurations are done.The slave repository is the Read only copy of the master and can only synchronize to the changes done in the master but cannot perform the changes itself.

I do not know exactly for what you are using teh master slave in your landscape.

But in most cases this concept has been designed to use it basically for backups and transport.

When you want to protect your repository from downtime failures or for transporting your repository data from one server to the other we use the master slave repository.

You can refer the below link to understand in detail teh working of Master Slave in MDM:

https://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/c06943e3-e881-2b10-ada0-ec469085...

Hope It Helped

Thanks & Regards

Simona Pinto

siowfong_chen
Contributor
0 Kudos

Thanks for the info Simon. Frankly speaking, I am still a bit hazy about the use of slave repository. By reading the functionalities, I thought I could use it to make changes to the repository without affecting the users who are accessing it (from SRM-MDM point of view). For example, if for some reasons I need to make a structure change to the repository, I can make the changes to the master repository and this would still be transparent to the user. And once the changes are confirmed, this can be sychronised to the slave repository. To do it for this purpose, does that mean that I will have to point the user to always use the slave repository and not the master? It seems that the system is not smart enough to automatically use the slave repository if the master is being changed (ie not loaded).

Appreciate clarification on this.

Cheers!

SF