[
https://issues.apache.org/jira/browse/IGNITE-14226?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Dmitriy Sorokin updated IGNITE-14226:
-------------------------------------
Description:
There should be tests of rebalance for both in-memory caches and persistent
caches.
In case of persistent caches, the tests also should check both modes of
rebalance: full and historical (asserts is needed that the proper mode of
rebalance was worked).
Basic scenario:
Testing of rebalance triggered by two event types: node join and node left the
topology (baseline change in persistent mode).
Extended scenario:
Node join or left during rebalance process.
Test parameters:
# Initial node count (derived from test context);
# Cache count - the count of caches to create and data preload;
# Cache entry count - the count of entries per cache to preload;
# Cache entry size - the size of each cache entry in bytes;
# Rebalance thread pool size - the value for
{{IgniteConfiguration#rebalanceThreadPoolSize}} property (see [configuring
rebalance thread pool
title|https://ignite.apache.org/docs/latest/data-rebalancing#configuring-rebalance-thread-pool]),
expected that greater value makes rebalance faster.
# Rebalance batch size - the value for
{{IgniteConfiguration#rebalanceBatchSize}} property (see [other rebalance
properties|https://ignite.apache.org/docs/latest/data-rebalancing#other-properties]),
expected that greater value makes rebalance faster on large data or
[throttling|https://ignite.apache.org/docs/latest/data-rebalancing#throttling]
enabled (rebalanceThrottling > 0).
# Rebalance throttle - the value for {{IgniteConfiguration#rebalanceThrottle}}
property (see [rebalance message
throttling|https://ignite.apache.org/docs/latest/data-rebalancing#throttling],
[other rebalance
properties|https://ignite.apache.org/docs/latest/data-rebalancing#other-properties]),
0 - throttling disabled, greater value makes rebalance slower.
Test result: time taken to rebalance process.
was:
There should be tests of rebalance for both in-memory caches and persistent
caches.
In case of persistent caches, the tests also should check both modes of
rebalance: full and historical (asserts is needed that the proper mode of
rebalance was worked).
Basic scenario:
Testing of rebalance triggered by two event types: node join and node left the
topology (baseline change in persistent mode).
Extended scenario:
Node join or left during rebalance process.
Test parameters:
# Initial node count (derived from test context);
# Cache count - the count of caches to create and data preload;
# Cache entry count - the count of entries per cache to preload;
# Cache entry size - the size of each cache entry in bytes;
# Rebalance thread pool size - the value for
IgniteConfiguration.rebalanceThreadPoolSize property (see [configuring
rebalance thread pool
title|https://ignite.apache.org/docs/latest/data-rebalancing#configuring-rebalance-thread-pool]),
expected that greater value makes rebalance faster.
# Rebalance batch size - the value for IgniteConfiguration.rebalanceBatchSize
property (see [other rebalance
properties|https://ignite.apache.org/docs/latest/data-rebalancing#other-properties]),
expected that greater value makes rebalance faster on large data or
[throttling|https://ignite.apache.org/docs/latest/data-rebalancing#throttling]
enabled (rebalanceThrottling > 0).
# Rebalance throttle - the value for IgniteConfiguration.rebalanceThrottle
property (see [rebalance message
throttling|https://ignite.apache.org/docs/latest/data-rebalancing#throttling],
[other rebalance
properties|https://ignite.apache.org/docs/latest/data-rebalancing#other-properties]),
0 - throttling disabled, greater value makes rebalance slower.
Test result: time taken to rebalance process.
> Ducktape tests of rebalance
> ---------------------------
>
> Key: IGNITE-14226
> URL: https://issues.apache.org/jira/browse/IGNITE-14226
> Project: Ignite
> Issue Type: Test
> Reporter: Dmitriy Sorokin
> Assignee: Dmitriy Sorokin
> Priority: Major
>
> There should be tests of rebalance for both in-memory caches and persistent
> caches.
> In case of persistent caches, the tests also should check both modes of
> rebalance: full and historical (asserts is needed that the proper mode of
> rebalance was worked).
> Basic scenario:
> Testing of rebalance triggered by two event types: node join and node left
> the topology (baseline change in persistent mode).
> Extended scenario:
> Node join or left during rebalance process.
> Test parameters:
> # Initial node count (derived from test context);
> # Cache count - the count of caches to create and data preload;
> # Cache entry count - the count of entries per cache to preload;
> # Cache entry size - the size of each cache entry in bytes;
> # Rebalance thread pool size - the value for
> {{IgniteConfiguration#rebalanceThreadPoolSize}} property (see [configuring
> rebalance thread pool
> title|https://ignite.apache.org/docs/latest/data-rebalancing#configuring-rebalance-thread-pool]),
> expected that greater value makes rebalance faster.
> # Rebalance batch size - the value for
> {{IgniteConfiguration#rebalanceBatchSize}} property (see [other rebalance
> properties|https://ignite.apache.org/docs/latest/data-rebalancing#other-properties]),
> expected that greater value makes rebalance faster on large data or
> [throttling|https://ignite.apache.org/docs/latest/data-rebalancing#throttling]
> enabled (rebalanceThrottling > 0).
> # Rebalance throttle - the value for
> {{IgniteConfiguration#rebalanceThrottle}} property (see [rebalance message
> throttling|https://ignite.apache.org/docs/latest/data-rebalancing#throttling],
> [other rebalance
> properties|https://ignite.apache.org/docs/latest/data-rebalancing#other-properties]),
> 0 - throttling disabled, greater value makes rebalance slower.
> Test result: time taken to rebalance process.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
