[
https://issues.apache.org/jira/browse/HDDS-8599?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17721863#comment-17721863
]
Siyao Meng commented on HDDS-8599:
----------------------------------
Hi Sumit, this was detailed in design doc section 6.2. There are 3 approaches
we could take and we had quite a few discussions. But we have not eventually
determined on which optimization route to take yet. [~ppogde] [~aswinshakil]
> [snapshot] performance impact for key & directory deleting service
> ------------------------------------------------------------------
>
> Key: HDDS-8599
> URL: https://issues.apache.org/jira/browse/HDDS-8599
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Sumit Agrawal
> Priority: Minor
>
> Scenario:
> # Huge number of keys and directories are created in bucket (FSO)
> # Snapshot is taken for the bucket
> # Recursive delete of keys and directory
> # keys and directories are added to deletedTable and deletedDirectoryTable
> KeyDeletingService:
> * iterator all keys in deleted Table
> ** if keys are present in snapshot, its skipped
> ** else added for blockGroupList for deletion
> DirectoryDeletingService:
> * iterator all directory in deleted directory table
> ** if directory is present in snapshot, its skipped
> ** else perform further action for deletion
> So if number of keys and directory are huge, it will keep iterating always to
> get candidate keys/directory for deletion.
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
