[
https://issues.apache.org/jira/browse/IGNITE-1093?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Anton Vinogradov updated IGNITE-1093:
-------------------------------------
Comment: was deleted
(was: I've created simple test to check how much time takes to relocated
40_000_000 entries:
{
Ignite ignite = startGrid(0);
try (IgniteDataStreamer<Integer, String> stmr =
ignite.dataStreamer(CACHE_NAME_DHT)) {
for (int i = 0; i < 40_000_000; i++)
stmr.addData(i, KEY_VAL);
}
startGrid(1);
stopAllGrids();
}
By default RebalanceTimeout equals to 10 seconds. In this case new node
requests entries each 10-20 seconds. And old node creates and sends message
with entries. Sounds bad, IMHO.
And in this case we have following metrics:
data load takes 81 s
suply message preparation takes 49 s
send-receive (from send to handle start) takes 399 s
put gained entries to cache takes 118 s
So, I set cacheCfg.setRebalanceTimeout(1000000) and gain:
data load takes 81s
suply message preparation takes 49 s
send-receive (from send to handle start) takes 45 s
put gained entries to cache takes 118 s
Seems that we have to change default value of rebalanceTimeout)
> Rebalancing with default parameters is very slow
> ------------------------------------------------
>
> Key: IGNITE-1093
> URL: https://issues.apache.org/jira/browse/IGNITE-1093
> Project: Ignite
> Issue Type: Bug
> Components: cache
> Affects Versions: sprint-7
> Reporter: Pavel Konstantinov
> Assignee: Anton Vinogradov
> Priority: Critical
> Fix For: sprint-8
>
>
> # Start one node with partitioned cache with one backup.
> # Load into the cache 40billions of keys using DataStreamer
> # Start second node on the same host
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
