You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@kudu.apache.org by "Todd Lipcon (JIRA)" <ji...@apache.org> on 2018/03/27 22:47:00 UTC

[jira] [Created] (KUDU-2383) Document and possibly track memory usage of maintenance threads

Todd Lipcon created KUDU-2383:
---------------------------------

             Summary: Document and possibly track memory usage of maintenance threads
                 Key: KUDU-2383
                 URL: https://issues.apache.org/jira/browse/KUDU-2383
             Project: Kudu
          Issue Type: Improvement
          Components: tserver
            Reporter: Todd Lipcon


Maintenance Manager threads use memory in both read-side and write-side block decoders and encoders. In a server running a workload with a table with ~250 string columns, I see each thread using between 100-200MB of RAM. If a user isn't aware of this memory usage they may under-provision the heap.

We should probably reserve some amount of memory for the MM threads, potentially track them via a MemTracker, and warn if the total configured heap (cache + MM threads estimate) is too close to the soft limit, leaving no room for MRS/DMS.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)