[jira] [Commented] (BOOKKEEPER-945) Add counters to track the activity of auditor and replication workers

["ASF GitHub Bot (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/BOOKKEEPER-945?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15456112#comment-15456112 ] 

ASF GitHub Bot commented on BOOKKEEPER-945:
-------------------------------------------

Github user rithin-shetty commented on the issue:

    https://github.com/apache/bookkeeper/pull/57
  
    The following tests have failed. These pass on my machine. I think these are flappers:
    
    testBookie(org.apache.bookkeeper.benchmark.TestBenchmark)  Time elapsed: 63.403 sec  <<< ERROR! java.lang.Exception: test timed out after 60000 milliseconds
    testReadThroughputLatency(org.apache.bookkeeper.benchmark.TestBenchmark)  Time elapsed: 64.472 sec  <<< ERROR! java.lang.Exception: test timed out after 60000 milliseconds



> Add counters to track the activity of auditor and replication workers
> ---------------------------------------------------------------------
>
>                 Key: BOOKKEEPER-945
>                 URL: https://issues.apache.org/jira/browse/BOOKKEEPER-945
>             Project: Bookkeeper
>          Issue Type: Improvement
>          Components: bookkeeper-server
>    Affects Versions: 4.5.0
>            Reporter: Rithin Shetty
>            Assignee: Rithin Shetty
>            Priority: Minor
>             Fix For: 4.5.0
>
>
> Once we enable auto recovery, auditor and replication workers start their activity. Today there is no way to monitor it using counters. This is a bug to track various activities of auditor and replication workers like: 
> - Time taken by auditor to build the bookie->ledger list 
> - No. of under replicated ledgers detected 
> - Time taken by auditor to publish the under replicated ledger list 
> - Time taken by auditor to check all the ledgers in the cluster 
> - No. of ledgers replicated by each replication worker 
> - No. of entries and bytes of data read and written by each replication worker
> - Auditor can also report the distribution of ledgers within the cluster: how many bookies own a piece of ledger, etc. 



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