[
https://issues.apache.org/jira/browse/HBASE-19441?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16280819#comment-16280819
]
Appy commented on HBASE-19441:
------------------------------
bq. Remember that backups are client driven (per some design review from a long
time ago), so queuing is tough to reason about (we have no "centralized"
execution system to use).
We have centralized execution system - procedureV2. Queueing is totally
possible.
We have good amount of time before 2.1, wouldn't it be reasonable to move
things to procedure framework. The benefits would be- Backup operation can
avoid using BackupMetaTable to maintain intermediate state (since procs have
support for state, WAL, recovery on crash, etc). That in turn may help get rid
of snapshot-restore of backuptable. Which means multiple backups can progress
in parallel.
> Implement retry logic around starting exclusive backup operation
> ----------------------------------------------------------------
>
> Key: HBASE-19441
> URL: https://issues.apache.org/jira/browse/HBASE-19441
> Project: HBase
> Issue Type: Improvement
> Components: backup&restore
> Reporter: Josh Elser
> Fix For: 3.0.0
>
>
> {quote}
> Specifically, the client does a checkAndPut to specifics coordinates in the
> backup table and throws an exception when that fails. Remember that backups
> are client driven (per some design review from a long time ago), so queuing
> is tough to reason about (we have no "centralized" execution system to use).
> At a glance, it seems pretty straightforward to add some retry/backoff
> semantics to BackupSystemTable#startBackupExclusiveOperation().
> {quote}
> While we are in a state in which backup operations cannot be executed in
> parallel, it would be nice to provide some retry logic + configuration. This
> would alleviate users from having to build this themselves.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
