You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Ilya Suntsov (JIRA)" <ji...@apache.org> on 2016/08/31 11:14:20 UTC

[jira] [Commented] (IGNITE-3474) Grid hung during load test with warning "Failed to wait for partition map exchange"

    [ https://issues.apache.org/jira/browse/IGNITE-3474?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15451944#comment-15451944 ] 

Ilya Suntsov commented on IGNITE-3474:
--------------------------------------

I reproduced this issue on ignite-1.7.0

> Grid hung during load test with warning "Failed to wait for partition map exchange"
> -----------------------------------------------------------------------------------
>
>                 Key: IGNITE-3474
>                 URL: https://issues.apache.org/jira/browse/IGNITE-3474
>             Project: Ignite
>          Issue Type: Bug
>    Affects Versions: 1.6
>            Reporter: Ksenia Rybakova
>            Assignee: Semen Boikov
>            Priority: Critical
>         Attachments: 091826_id0_192.168.1.21_part.log, ignite-base-load-config.xml, run-load.properties, run-load.xml
>
>
> During atomic operations load test grid hung:
> {noformat}
> [09:22:17,688][WARN ][exchange-worker-#137%null%][GridCachePartitionExchangeManager] Failed to wait for partition map exchange [topVer=AffinityTopologyVersion [topVer=97, minorTopVer=0], node=0fe0c0f2-65e1-4ad5-a8c9-92db997fcb7d]. Dumpin
> g pending objects that might be the cause:
> ...
> {noformat}
> See attached file with more info from logs.
> Notes:
> 1) The issue was also reproduced with IGNITE-3239 fixed. So, IGNITE-3239 doesn't seem to be a root cause. 
> 2) The issue was also reproduced with other configs (not only atomic operations). Actually, most of laod tests failed due to this issue.
> Load test config:
> - Benchmark name: IgniteCacheRandomOperationBenchmark
> - Operations: put, put_all, get, get_all, invoke, invoke_all, remove, remove_all, put_if_absent, replace
> - Heap: 2Gb for servers, 1Gb for clients
> - 48 clients, 48 servers, 6 hosts
> - Preload amount: 2.5M
> - Key range: 5M
> - Number of caches: 12
> - Types of caches (atomicity mode): only atomic
> - Types of caches (tiered storage mode): different (onheap without eviction, onheap with eviction, offheap_tired, offheap_values)
> - Types of caches (indexing): different (with and without indexes)
> - Types of caches (cache mode): partitioned
> - Backups count: 1
> Attachments:
> - part of log with the problem (for one of the servers)
> - xml configs and property file that were used to run yardstick



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)