[
https://issues.apache.org/jira/browse/HDDS-9135?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hemant Kumar updated HDDS-9135:
-------------------------------
Description:
To fix [HDDS-9126|https://issues.apache.org/jira/browse/HDDS-9126], we added
and used *isSstFiltered* flag to achieve synchronization between snapshot
deletion and SST filtering service.
First of all, there is no need to achieve to synchronization in theory and we
should just use active snapshot check because in worst case scenario SST
filtering service will fail for the snapshot and in next snapshot run it will
be skipped because snapshot is deleted.
If the above doesn't work and it ends up in memory leaking or inconsistence DB
state, we should use locking for snapshot background services (not the same
lock used for snapshot APIs) instead of using *isSstFiltered*. Same locking
mechanism can be use to synchronize new background service related to snapshot
gets added.
was:
To fix [HDDS-9126|https://issues.apache.org/jira/browse/HDDS-9126], we added
and used *isSstFiltered* flag to achieve synchronization between snapshot
deletion and SST filtering service.
First of all, there is no need to achieve to synchronization in theory and we
should just use active snapshot check because in worst case scenario SST
filtering service will fail for the snapshot and in next snapshot run it will
be skipped because snapshot is deleted.
If the above doesn't work and it ends up in memory leaking or inconsistence DB
state, we should use locking for snapshot background services (not the same
lock we used to snapshot APIs) instead of using *isSstFiltered*. Same locking
mechanism can be use to synchronize new background service related to snapshot
gets added.
> Snapshot deletion and SST filtering service synchronization
> -----------------------------------------------------------
>
> Key: HDDS-9135
> URL: https://issues.apache.org/jira/browse/HDDS-9135
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Hemant Kumar
> Priority: Major
>
> To fix [HDDS-9126|https://issues.apache.org/jira/browse/HDDS-9126], we added
> and used *isSstFiltered* flag to achieve synchronization between snapshot
> deletion and SST filtering service.
> First of all, there is no need to achieve to synchronization in theory and we
> should just use active snapshot check because in worst case scenario SST
> filtering service will fail for the snapshot and in next snapshot run it will
> be skipped because snapshot is deleted.
> If the above doesn't work and it ends up in memory leaking or inconsistence
> DB state, we should use locking for snapshot background services (not the
> same lock used for snapshot APIs) instead of using *isSstFiltered*. Same
> locking mechanism can be use to synchronize new background service related to
> snapshot gets added.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
