jackrabbit-oak-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Vikas Saurabh (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (OAK-4623) Log more information when null DocumentNodeState is read for a child while fetching children
Date Sun, 21 Aug 2016 00:01:40 GMT

    [ https://issues.apache.org/jira/browse/OAK-4623?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15429539#comment-15429539
] 

Vikas Saurabh commented on OAK-4623:
------------------------------------

[~mreutegg], I agree with logging cached (or fresh as the case may be) document for the child
- but I'm feeling a little uneasy for forcing a read with maxCacheAge=0 as it'd update doc
cache just for logging purposes. I mean, I feel ok with invalidating a potential bad cache
entry - but I won't want to read it from remote just for logging. But, then I can also see
the counter-argument that this situation must never occur - so a penalty to read remote for
logging should be ok.

Also, I wonder if we should do reads in try{} to handle DocStore#find throwing DocStoreException.

> Log more information when null DocumentNodeState is read for a child while fetching children
> --------------------------------------------------------------------------------------------
>
>                 Key: OAK-4623
>                 URL: https://issues.apache.org/jira/browse/OAK-4623
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: documentmk
>            Reporter: Vikas Saurabh
>            Assignee: Vikas Saurabh
>            Priority: Minor
>             Fix For: 1.6
>
>
> {{DocumentNodeStore#getChildNodes}} validates node states of children in the returned
iterator and throws exception if a child is read as {{null}}.
> This is clearly an unexpected case (although, it has been observed in running systems).
The more probable cause for such a scenario would be to have inconsistent value in {{nodeChildrenCache}}.
It'd be useful to log more information about cache state to help investigating such cases.



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

Mime
View raw message