[
https://issues.apache.org/jira/browse/HDDS-5941?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Bharat Viswanadham updated HDDS-5941:
-------------------------------------
Description:
In SCM non-HA when we use rocksdb put, it should write to DB with sync turned
on. As in non-HA when power failure/system crashes, the operations which are
acknowledged by SCM might be lost in this kind of scenario. (As in rocks DB
with Sync false, the flush is asynchronous and it will not persist to storage
system)
In HA, this is not a problem because the guarantee is provided by ratis and
ratis logs.
was:
In OM non-HA when double buffer flushes, it should commit with sync turned on.
As in non-HA when power failure/system crashes, the operations which are
acknowledged by OM might be lost in this kind of scenario. (As in rocks DB with
Sync false, the flush is asynchronous and it will not persist to storage system)
In HA, this is not a problem because the guarantee is provided by ratis and
ratis logs.
> Enable sync option for SCM non-HA
> ----------------------------------
>
> Key: HDDS-5941
> URL: https://issues.apache.org/jira/browse/HDDS-5941
> Project: Apache Ozone
> Issue Type: Task
> Reporter: Bharat Viswanadham
> Assignee: Bharat Viswanadham
> Priority: Major
> Fix For: 0.5.0
>
>
> In SCM non-HA when we use rocksdb put, it should write to DB with sync turned
> on. As in non-HA when power failure/system crashes, the operations which are
> acknowledged by SCM might be lost in this kind of scenario. (As in rocks DB
> with Sync false, the flush is asynchronous and it will not persist to storage
> system)
>
> In HA, this is not a problem because the guarantee is provided by ratis and
> ratis logs.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
