jackrabbit-oak-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael Dürig (JIRA) <j...@apache.org>
Subject [jira] [Reopened] (OAK-4873) Avoid running GC too frequently
Date Mon, 19 Dec 2016 10:50:59 GMT

     [ https://issues.apache.org/jira/browse/OAK-4873?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Michael Dürig reopened OAK-4873:
--------------------------------

Reopening this as tests have shown users are too creative in working around the estimator.
That is, you can still run gc, upload a couple of GB content and run gc again seconds later,
which bears the risk of running into a transient {{SNFE}}. 

I think we should add a safeguard to {{SegmentRevisionGCMBean.startRevisionGC()}} so it would
not run if invoked more than once in 12 hours. Instead it would just log a warning and reflect
this fact via e.g. {{SegmentRevisionGCMBean.getStatus()}}. For internal purposes the 12 hours
defaults should be configurable via a feature flag. 

> Avoid running GC too frequently
> -------------------------------
>
>                 Key: OAK-4873
>                 URL: https://issues.apache.org/jira/browse/OAK-4873
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: segment-tar
>    Affects Versions: Segment Tar 0.0.16
>            Reporter: Michael Dürig
>            Assignee: Michael Dürig
>              Labels: compaction, gc
>             Fix For: 1.5.17, 1.6
>
>
> As {{RevisionGCMBean.startRevisionGC()}} can be used to manually invoke a gc cycle, there
is the danger of running into a {{SNFE}} when gc is run multiple times in quick succession
(due to the retention time being based on number of generations). We should come up with a
mechanism to prevent this scenario. 



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

Mime
View raw message