[ 
https://issues.apache.org/jira/browse/SOLR-8709?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15226550#comment-15226550
 ] 

Joel Bernstein commented on SOLR-8709:
--------------------------------------

Commit settings were:

1 second hard commits
2 second soft commits

The softCommit is what matters because the TopicStream is working off of index 
snapshots.

In reviewing the code, it certainly looks like there is room for out-of-order 
version numbers to cross commit boundaries, but I would expect to see it happen 
in a stress test.

I'm thinking about changing this ticket to maintain a checksum of the version 
numbers for a time window. Then resending the time window if the checksums 
don't match. This is much more manageable then attempting to track all version 
numbers for a time window. This won't provide guaranteed *one time* delivery of 
documents, but it will provide guaranteed delivery of all documents in a topic, 
with a reasonable expectation of one time delivery.

> Account for out-of-order version numbers in the TopicStream
> -----------------------------------------------------------
>
>                 Key: SOLR-8709
>                 URL: https://issues.apache.org/jira/browse/SOLR-8709
>             Project: Solr
>          Issue Type: Bug
>            Reporter: Joel Bernstein
>
> Currently the TopicStream can miss documents if version numbers are received 
> out-of-order. The TopicStream sorts on version number so it will only miss 
> out-of-order versions that span commit boundaries.
> In order to resolve this issue we can adopt an approach that keeps a set of 
> the last N version numbers sent for each Topic.  As the documents are scanned 
> we can check for documents within this time window that do not appear in the 
> sent set. These documents can then be sent.



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

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to