How data region is propagated to other nodes

classic Classic list List threaded Threaded
6 messages Options
dhiman_nikhil dhiman_nikhil
Reply | Threaded
Open this post in threaded view
|

How data region is propagated to other nodes

Server level config:
Server: Node 1  
CacheKey: "location"
DataRegion: "location"
persistence: true

Server: Node 2
CacheKey: "category"
DataRegion: "category"
persistence : false

Both belong to same cluster.
Since Data is partitioned across nodes when I start Node 1 and then Node 2,
will "location" region is also created in Node 2 since data will be
partitioned for CacheKey: "location"  ??



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
ezhuravlev ezhuravlev
Reply | Threaded
Open this post in threaded view
|

Re: How data region is propagated to other nodes

Hi,

No, Data region configuration won't be propagated to other nodes in the cluster. DataRegion is only a local configuration, so, you have to configure it on each node. Moreover, these the same data region on different nodes can have different maxSize, which could be useful in case of using NodeFilter for caches.

Best Regards,
Evgenii

чт, 2 мая 2019 г. в 14:01, dhiman_nikhil <[hidden email]>:
Server level config:
Server: Node 1 
CacheKey: "location"
DataRegion: "location"
persistence: true

Server: Node 2
CacheKey: "category"
DataRegion: "category"
persistence : false

Both belong to same cluster.
Since Data is partitioned across nodes when I start Node 1 and then Node 2,
will "location" region is also created in Node 2 since data will be
partitioned for CacheKey: "location"  ??



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
dhiman_nikhil dhiman_nikhil
Reply | Threaded
Open this post in threaded view
|

Re: How data region is propagated to other nodes

This post was updated on .
Thanks, Evgenii

As per documentation, I can do persistence in data region only. If I wanted a
cache key to be persisted then I have to make sure that data region to be
replicated across all nodes.

Basically, my requirement is to give multi-tenancy support[Multiple
Applications can communicate to Ignite Cluster]. To prevent conflict of
CacheKey, I will be writing an Ignite Client[Wrapper over IgniteClient
library] and passing on the limiting functionality to Services[Like they can
get CacheKey, but not been able to create it].

If a new Application comes and demands a CacheKey with persistence ON[dedicated unique Data Region], we
should be able to provide it without reloading the whole nodes of Ignite
cluster.

Regards,
Nikhil

--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
ezhuravlev ezhuravlev
Reply | Threaded
Open this post in threaded view
|

Re: How data region is propagated to other nodes

Then you can simply create 2 different data regions for all server nodes in the cluster - one with persistence and one without. After that, you can decide, how you want to store your key - with persistence or without.

Evgenii

пт, 3 мая 2019 г. в 08:27, dhiman_nikhil <[hidden email]>:
As per documentation, I can do persistence in data region only. If I wanted a
cache key to be persisted then I have to make sure that data region to be
replicated across all nodes.

Basically, my requirement is to give multi-tenancy support[Multiple
Applications can communicate to Ignite Cluster]. To prevent conflict of
CacheKey, I will be writing an Ignite Client[Wrapper over IgniteClient
library] and passing on the limiting functionality to Services[Like they can
get CacheKey, but not been able to create it].

If a new Application comes and demands a CacheKey with persistence ON, we
should be able to provide it without reloading the whole nodes of Ignite
cluster.



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
dhiman_nikhil dhiman_nikhil
Reply | Threaded
Open this post in threaded view
|

Re: How data region is propagated to other nodes

Can I create a data region dynamically on every node in the cluster?



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
ezhuravlev ezhuravlev
Reply | Threaded
Open this post in threaded view
|

Re: How data region is propagated to other nodes

In reply to this post by ezhuravlev
Data regions can't be created dynamically, because it's needed to preallocate memory. I believe that in your use case you can just create it statically on cluster start, is there a problem with that?

Evgenii

пт, 3 мая 2019 г. в 12:54, dhiman_nikhil <[hidden email]>:
Can I create data regions on All Servers dynamically?



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/