[
https://issues.apache.org/jira/browse/IGNITE-1152?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14802893#comment-14802893
]
Anton Vinogradov commented on IGNITE-1152:
------------------------------------------
doing
for (int i = 0; i < 100000; i++) {
n1.cache("c1").put(i, 1);
n2.cache("c1").put(i, 1);
n3.cache("c1").put(i, 1);
n4.cache("c1").put(i, 1);
}
before counting backups solves problem.
Seems backup evictions depends on cache operations.
Result before warmup:
<16:16:34,071><DEBUG><main> Node n1: : pri=238, bak=341
<16:16:34,072><DEBUG><main> Node n2: : pri=230, bak=633
<16:16:34,072><DEBUG><main> Node n3: : pri=288, bak=410
<16:16:34,073><DEBUG><main> Node n4: : pri=268, bak=260
Result after warmup:
<16:15:55,716><DEBUG><main> Node n1: : pri=238, bak=340
<16:15:55,717><DEBUG><main> Node n2: : pri=230, bak=349
<16:15:55,717><DEBUG><main> Node n3: : pri=288, bak=293
<16:15:55,717><DEBUG><main> Node n4: : pri=268, bak=260
> Distribution of backup partitions is not uniform
> ------------------------------------------------
>
> Key: IGNITE-1152
> URL: https://issues.apache.org/jira/browse/IGNITE-1152
> Project: Ignite
> Issue Type: Bug
> Components: cache
> Affects Versions: 1.1.4
> Reporter: Alexey Kuznetsov
> Assignee: Alexey Goncharuk
> Attachments: CacheBackupPartitionsTest.java
>
>
> I started 4 nodes with partitioned cache with 1 backup.
> And found that primary parts more or less uniform, but backup parts - not:
> Node n1: : pri=244, bak=367
> Node n2: : pri=260, bak=590
> Node n3: : pri=244, bak=367
> Node n4: : pri=260, bak=590
> Code to test this issue attached.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
