[
https://issues.apache.org/jira/browse/OAK-12040?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18046172#comment-18046172
]
Julian Sedding commented on OAK-12040:
--------------------------------------
In order to assess the impact of the change, I ran the {{SegmentStoreMigrator}}
tool to migrate a {{{}TarPersistence to an {{AzurePersistence}}{}}}. The
TAR-file based segmentstore consisted of 10 archives containing ~67k segments,
with a total size of ~1.85GB. The choice of the {{TarPersistence}} as the
source is to remove read I/O from the equation and create a scenario that is
dominated entirely by write I/O.
Running the migration tookÂ
* before the change: 14min 3sec
* after the change: 7min 49sec
That is a ~45% speedup.
> segment-azure: reduce HTTP requests for writes
> ----------------------------------------------
>
> Key: OAK-12040
> URL: https://issues.apache.org/jira/browse/OAK-12040
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: segment-azure
> Reporter: Julian Sedding
> Assignee: Julian Sedding
> Priority: Minor
>
> Currently,
> {{org.apache.jackrabbit.oak.segment.azure.AzureSegmentArchiveWriter#doWriteArchiveEntry}}
> issues two http requests for every write. One to upload the blob and a
> second one to set the metadata on the blob. This can be optimized to use a
> single HTTP request instead.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
