[
https://issues.apache.org/jira/browse/ARTEMIS-1700?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16375504#comment-16375504
]
Francesco Nigro commented on ARTEMIS-1700:
------------------------------------------
Thanks!
Just a couple of tips on it.
Considering how the TimedBuffer works, if you need scalability vs a big
number of writers with blocking API like NIO/MAPPED you need to:
- run ./artemis perf-journal --journal-type MAPPED --sync --verbose
--sync-writes
- read the histogram and use at least the latencies starting from the 0.90
percentiles, then add your network latency and the results would be the
optimal buffer-timeout.
As a simpler rule of thumb I use about twice the default timeout proposed
by Artemis: you will loose something on the single producer/consumer case
while gaining more chances of batched writes with multiple
producers/consumers.
While, if you are just benchmarking and do not care about power failures
but just OS failures try buffer-timeout = 0 and datasync = false on
broker.xml with MAPPED.
Il giorno sab 24 feb 2018 alle ore 02:51 Qihong Xu (JIRA) <j...@apache.org>
> Server stopped responding and killed itself while exiting paging state
> ----------------------------------------------------------------------
>
> Key: ARTEMIS-1700
> URL: https://issues.apache.org/jira/browse/ARTEMIS-1700
> Project: ActiveMQ Artemis
> Issue Type: Bug
> Components: Broker
> Affects Versions: 2.4.0
> Reporter: Qihong Xu
> Priority: Major
> Attachments: artemis.log
>
>
> We are currently experiencing this error while running stress test on artemis.
>
> Basic configuration:
> 1 broker ,1 topic, pub-sub mode.
> Journal type = MAPPED.
> Threadpool max size = 60.
>
> In order to test the throughput of artemis we use 300 producers and 300
> consumers. However we found that sometimes when artemis exit paging state, it
> will stop responding and kill itself. This situatuion happened on some
> specific servers.
>
> Details can be found in attached dump file.
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
