[
https://issues.apache.org/jira/browse/OAK-11132?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Nuno Santos resolved OAK-11132.
-------------------------------
Fix Version/s: 1.70.0
Resolution: Done
> AOT blob downloader: may download a blob for a node that has already been
> indexed
> ---------------------------------------------------------------------------------
>
> Key: OAK-11132
> URL: https://issues.apache.org/jira/browse/OAK-11132
> Project: Jackrabbit Oak
> Issue Type: Bug
> Components: indexing
> Reporter: Nuno Santos
> Priority: Minor
> Fix For: 1.70.0
>
>
> The AOT downloader tries to stay ahead of the indexer, as there is no need to
> download blobs for nodes that were already indexed. When reading a line in
> the FFS, it checks if the position of this line is ahead of the current
> indexing position, and if not, it skips downloading any blob that this line
> may have. However, this check is done as soon as the line is read from the
> FFS. Before the blob is actually downloaded, it the request to downloaded is
> enqueued and must be picked up by a downloader thread. By this time, the
> indexer may have caught up.
> So we should check again just before downloading the blob.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
