[jira] [Commented] (OAK-3399) 5sec retry loop before declaring lease failure (was: as a last resort try to update it there)

["Stefan Egli (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/OAK-3399?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14743418#comment-14743418 ] 

Stefan Egli commented on OAK-3399:
----------------------------------

alternatively - as a less intrusive solution: do a 5 x 1sec retry loop when detecting a lease failure before actually declaring it. That would delay the failure-action by 5sec (which I see as non-problematic - it would anyway have been done), but in the best case the lease-update-thread just had a hickup and it was lucky enough to update it again - in which case we don't have to stop. This would probably help in development as it supports the laptop-wake-from-sleep case..

> 5sec retry loop before declaring lease failure (was: as a last resort try to update it there)
> ---------------------------------------------------------------------------------------------
>
>                 Key: OAK-3399
>                 URL: https://issues.apache.org/jira/browse/OAK-3399
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: core
>    Affects Versions: 1.3.6
>            Reporter: Stefan Egli
>
> Not sure if this is a good idea - but when the lease check finds out that the lease is valid for less than the margin (which is 20sec by default), it could then try to update the lease directly. The 'ugly' part of this is that normally this job is done by a dedicated background 'lease update' thread - and now it would be done in the context of a user-thread 'very much under the hood'. but the advantage would be that it would eg catch the 'laptop-sleep' situation when running a single vm ..



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