[
https://issues.apache.org/jira/browse/IGNITE-1093?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14701581#comment-14701581
]
Anton Vinogradov commented on IGNITE-1093:
------------------------------------------
I runned benchmark at new implementation:
3 server nodes, 1 client node.
1M <Int,Int> entries placed to cache.
Client node permanently gets and puts data to cache.
with params:
rebalancing uses 4 threads and pregenerate 3 batches of regular size (512K)
with options:
JVM_OPTS=${JVM_OPTS}" -DIGNITE_QUIET=false -Xms5g -Xmx5g \
-Xloggc:./gc.log \
-XX:+PrintGCDetails \
-verbose:gc \
-XX:+UseParNewGC \
-XX:+UseConcMarkSweepGC \
-XX:+UseTLAB \
-XX:NewSize=128m \
-XX:MaxNewSize=128m \
-XX:MaxTenuringThreshold=0 \
-XX:SurvivorRatio=1024 \
-XX:+UseCMSInitiatingOccupancyOnly \
-XX:CMSInitiatingOccupancyFraction=60 \
"
Waited 15 seconds from start before chart recording. Entries succesfully placed
to cache.
And waited 10 secons before stop first server node and 15 more seconds before
start one more.
Results can be found at rebalancing.zip.
> 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: ignite-1.4
>
>
> # 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)
