[ 
https://issues.apache.org/jira/browse/OAK-3054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14973653#comment-14973653
 ] 

Chetan Mehrotra commented on OAK-3054:
--------------------------------------

Approach looks fine. Some comments
# Rename {{getExecutionNodesCount}} ->  {{getIndexedNodesCount}}
# {{latestErrorWarn}} - This should probably be reset when a warn log is done
# {{latestError}}, {{latestErrorTime}} - Should be reset in fixed
# You can use {{com.google.common.base.Throwables#getStackTraceAsString}} 
instead of creating a custom utility method
# Probably mark {{failing}} as volatile

Also was thinking that we should catch {{Exception}} instead of 
{{CommitFailedException}} as in many cases we might not be wrapping exception 
occurring in various {{IndexEditors}}. Many times I have seen exception being 
logged by Sling scheduler as it gets propagated 

> IndexStatsMBean should provide some details if the async indexing is failing
> ----------------------------------------------------------------------------
>
>                 Key: OAK-3054
>                 URL: https://issues.apache.org/jira/browse/OAK-3054
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: query
>            Reporter: Chetan Mehrotra
>            Assignee: Alex Parvulescu
>            Priority: Minor
>              Labels: resilience, tooling
>             Fix For: 1.3.9, 1.0.23, 1.2.8
>
>         Attachments: OAK-3054-v1.patch
>
>
> If the background indexing fails for some reason it logs the exception for 
> the first time then it logs the exception like _The index update failed ..._. 
> After that if indexing continues to fail then no further logging is done so 
> as to avoid creating noise.
> This poses a problem on long running system where original exception might 
> not be noticed and index does not show updated result. For such cases we 
> should expose the indexing health as part of {{IndexStatsMBean}}. Also we can 
> provide the last recorded exception. 
> Administrator can then check for MBean and enable debug logs for further 
> troubleshooting



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

Reply via email to