[
https://issues.apache.org/jira/browse/IGNITE-28279?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18107718#comment-18107718
]
Ignite TC Bot commented on IGNITE-28279:
----------------------------------------
Checked the duplicate resolution and related Ignite 2 context. This duplicate
has no separate documentation deliverable; any required documentation should
stay on the original implementation/documentation chain. Removing only Docs
Required on the duplicate only; Release Notes Required is left unchanged.
> Snapshot create operation should be allowed during Rolling Upgrade process
> --------------------------------------------------------------------------
>
> Key: IGNITE-28279
> URL: https://issues.apache.org/jira/browse/IGNITE-28279
> Project: Ignite
> Issue Type: Improvement
> Reporter: Sergey Chugunov
> Assignee: Aleksandr Chesnokov
> Priority: Major
> Labels: IEP-132, ise
> Fix For: 2.19
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> Ticket IGNITE-26991 disabled all distributed operations when Rolling Upgrade
> is in progress. It is a safe solution aiming to protect users from tricky
> corner cases but it brings other corner cases with it.
> One could imagine a situation when Rolling Upgrade procedure takes more time
> than initially planned and overlapse with a cron task for creating a
> snapshot. In that case cron task will fail and force the user to create the
> snapshot manually or skip it which is also not good.
> At the same time snapshot creation operation is RU-safe because at the end of
> the day it copies data files that are backward compatible. Even if a snapshot
> was created on an old node, it can be restored on a new one safely (but not
> wise-versa).
> Thus we need to come up with a more flexible mechanism allowing to execute
> some distributed operations on a cluster in the middle of RU process.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
