[jira] [Commented] (KUDU-1366) Consider switching to jemalloc

["Todd Lipcon (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/KUDU-1366?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15727705#comment-15727705 ] 

Todd Lipcon commented on KUDU-1366:
-----------------------------------

I ran another experiment yesterday with jemalloc on a cluster that was showing a lot of threads blocked in tcmalloc locks. The workload in question has a lot of tablets per server and extreme load from Impala. With jemalloc the blocking basically went away and I saw fewer RPC timeouts, etc. Just another data point that it's probably a better choice for us.

> Consider switching to jemalloc
> ------------------------------
>
>                 Key: KUDU-1366
>                 URL: https://issues.apache.org/jira/browse/KUDU-1366
>             Project: Kudu
>          Issue Type: Bug
>          Components: perf
>            Reporter: Todd Lipcon
>         Attachments: Kudu Benchmarks.pdf, Kudu Benchmarks.pdf
>
>
> We spend a fair amount of time in the allocator. While we could spend some time trying to use arenas more, it's also worth considering switching allocators. I ran a few quick tests with jemalloc 4.1 and it seems like it might be better than the version of tcmalloc that we use (and has much more active development)



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