off heap and jvm tuning

classic Classic list List threaded Threaded
2 messages Options
neerajbhatt neerajbhatt
Reply | Threaded
Open this post in threaded view
|

off heap and jvm tuning

Hello All

As per my understanding by default ignite 2.0 keeps all objects in off heap by default
We have kept java heap as 8 GB for all machines in cluster

1. When we add or remove a node (Partitioned cached, with backup one) we see that heap usage in one machine of cluster accordingly decreased and increased.

Ideally if off heap is used, heap usage should not fluctuate when we add/remove nodes in cluster

thanks


afedotov afedotov
Reply | Threaded
Open this post in threaded view
|

Re: off heap and jvm tuning

Hello.

Offheap mode mostly refers to the way the cache entries are stored.
There are many other parts of the product that rely on the Java heap like any Java application.
In case of a node joining/leaving a topology, partition rebalancing process will take place, which
inevitably will affect heap.

Kind regards,
Alex.

On Mon, Jun 19, 2017 at 12:21 PM, neerajbhatt [via Apache Ignite Users] <[hidden email]> wrote:
Hello All

As per my understanding by default ignite 2.0 keeps all objects in off heap by default
We have kept java heap as 8 GB for all machines in cluster

1. When we add or remove a node (Partitioned cached, with backup one) we see that heap usage in one machine of cluster accordingly decreased and increased.

Ideally if off heap is used, heap usage should not fluctuate when we add/remove nodes in cluster

thanks





If you reply to this email, your message will be added to the discussion below:
http://apache-ignite-users.70518.x6.nabble.com/off-heap-and-jvm-tuning-tp13950.html
To start a new topic under Apache Ignite Users, email [hidden email]
To unsubscribe from Apache Ignite Users, click here.
NAML