Welcome to the new Gigaspaces XAP forum. To recover your account, please follow these instructions.

Ask Your Question
0

Can XAP repartition spaces defined with Processing Units

Suppose we deploy a processing unit defining a space configured to be split in 2 partitions with 1 backup each.

Can we change this configuration at runtime? Add or remove partitions and have data transfered to the new nodes?

To be clear, I'm talking about PU deployment, not IMDG.

Thnx

{quote}This thread was imported from the previous forum. For your reference, the original is [available here|http://forum.openspaces.org/thread.jspa?threadID=3521]{quote}

asked 2010-10-30 18:53:31 -0500

sweldim gravatar image

updated 2013-08-08 09:52:00 -0500

jaissefsfex gravatar image
edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted
0

See the Rebalancing Utility under this section:
http://www.gigaspaces.com/wiki/display/SBP/MovingintoProductionChecklist#MovingintoProductionChecklist-Rebalancing
It supports re-balancing of a PU with an embedded space. This will allow you to expand a running PU with an embedded space while the system is running and stretch it across newly started GSCs nodes.
Shay

answered 2010-10-30 22:42:50 -0500

shay hassidim gravatar image
edit flag offensive delete link more

Comments

I run some tests and -as far as I understood-, re-balancing is about relocating primary or backup partitions to other GSCs/Hosts.

This process cannot change the number of partitions.

Unless you start with of configuration of, lets say, 1000 partitions located on 4 GSCs and then expand to 1000 GSCs, rebalancing does not provide a means to expand and shrink at runtime.

sweldim gravatar imagesweldim ( 2010-10-31 16:33:00 -0500 )edit

The "re-balancing" approach is the approach every data grid product is using. Some of the products simply have a default number of partitions used so this whole issue is hidden from most of the users. GigaSpaces allows you to specify this at the deployment time. This provides you better granularity how the system will "slice" the data across the deployed nodes.

You don't have to start with 1000 partitions to use re-balancing utility. You should use the "expansion factor" as the factor you should use to calculate the amount of partitions you should initially deploy with.
The expansion factor will be the amount of machines (or GSCs) your system might have in the future before you will need to restart the system and increase the number of partitions. You can start with 20 partitions with 2 GSC across 2 machines and later expand this to have the same 20 partitions running across 20 machines. This will allow you to increase your data grid capacity in an order of magnitude (x10).

Shay

shay hassidim gravatar imageshay hassidim ( 2010-11-01 07:56:45 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower

Stats

Asked: 2010-10-30 18:53:31 -0500

Seen: 104 times

Last updated: Oct 30 '10