[
https://issues.apache.org/jira/browse/HBASE-18073?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16017782#comment-16017782
]
Josh Elser commented on HBASE-18073:
------------------------------------
This is strange
{noformat}
2017-05-19 14:12:36,337 TRACE
[hw10447.local,16201,1495216952120_ChoreService_2] hbase.ChoreService:
onChoreMissedStartTime
2017-05-19 14:12:36,337 TRACE
[hw10447.local,16201,1495216952120_ChoreService_2] hbase.ChoreService: Chore
name: FileSystemUtilizationChore
2017-05-19 14:12:36,338 TRACE
[hw10447.local,16201,1495216952120_ChoreService_2] hbase.ChoreService: Chore
period: 30000
2017-05-19 14:12:36,338 TRACE
[hw10447.local,16201,1495216952120_ChoreService_2] hbase.ChoreService: Chore
timeBetweenRuns: 60000
2017-05-19 14:12:36,338 INFO
[hw10447.local,16201,1495216952120_ChoreService_2]
quotas.FileSystemUtilizationChore: Chore: FileSystemUtilizationChore missed its
start time
{noformat}
Even though the period is 30s, the thread is being invoked at 60s. So, every
time it wakes up, it thinks it needs to run (and rightfully trigger the
warning). Maybe this is just a silly bug I'm not seeing yet in this space quota
chore. Will continue to poke.
> ScheduledChore with delay longer than period never runs
> -------------------------------------------------------
>
> Key: HBASE-18073
> URL: https://issues.apache.org/jira/browse/HBASE-18073
> Project: HBase
> Issue Type: Improvement
> Reporter: Josh Elser
> Assignee: Josh Elser
> Priority: Minor
>
> (Obligatory: saw this on a fork -- need to confirm if this affects the Apache
> branches, and which branches if so)
> If a ScheduledChore is configured with a delay that is longer than the
> period, the Chore never actually gets run, instead repeatedly complains that
> the Chore missed its start time.
> {noformat}
> 2017-05-18 17:17:06,606 TRACE [server.com,16020,1495125783052_ChoreService_1]
> hbase.ChoreService: onChoreMissedStartTime
> 2017-05-18 17:17:06,612 TRACE [server.com,16020,1495125783052_ChoreService_1]
> hbase.ChoreService: Chore name: FileSystemUtilizationChore
> 2017-05-18 17:17:06,612 TRACE [server.com,16020,1495125783052_ChoreService_1]
> hbase.ChoreService: Chore period: 30000
> 2017-05-18 17:17:06,612 TRACE [server.com,16020,1495125783052_ChoreService_1]
> hbase.ChoreService: Chore timeBetweenRuns: 60000
> 2017-05-18 17:17:06,612 INFO [server.com,16020,1495125783052_ChoreService_1]
> quotas.FileSystemUtilizationChore: Chore: FileSystemUtilizationChore missed
> its start time
> {noformat}
> It seems like this might be an edge-case for the first invocation of the
> chore. Need to read the code closer.
> The workaround is to just ensure that the delay is always less than the
> period.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
