[
https://issues.apache.org/jira/browse/IGNITE-12968?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17198369#comment-17198369
]
Maxim Muzafarov commented on IGNITE-12968:
------------------------------------------
[~dmagda]
Thanks for such a detailed response. I'm not against rearranging topics itself
and document simplification but some of the important details were cleaned and
technical mistakes were added.
Here are my points:
# {{recovery procedure}} from my point of view mostly related to crash-recovery
Ignite machinery and it's better to use {{resotre}} instead
# the link to the {{ Ignite Persistence storage directory}} were added without
mentioning {{binary_meta}} and {{marshaller}} folders with is also important
for the snapshot structure
# restore procedure topic - checkpoint and wal directories must also be cleaned
prior to recovery since it can trigger unnecessary recovery mechanisms at a
node startup
# I think it's important to mention that partition backups are also included in
a snapshot (this was removed)
# I think it's important to mention that native persistence work directory will
require additional space to perform snapshot procedures (up to 1x) (this was
removed)
# joining a new node doesn't interrupt the snapshot procedure, only the
baseline node left interrupts it.
# the term `incremental snapshots` hasn't been introduced for users that are
not familiar with enterprise software of some vendors, so I think it has no
sense here. I think It will be better to provide some links with the Ignite
snapshots and other vendors snapshot functionality comparison on some external
resources like DZone etc.
This is a brief look over inaccuracies I've found during reading the
documentation page.
Denis,
I think we should apply the same rules of the review process of new
documentation pages as we do for the other Ignite source code. From my point of
view, we should fix the original PR and merge it when the review is finished
rather than merging incomplete changes and fix them in the master branch.
And personally my thing, it's a very huge demotivation for me for writing new
documentation pages when I see that my PR changed more than 80% in the master
branch after merge. It really hurts me :-)
> Create cluster snapshot documentation pages
> -------------------------------------------
>
> Key: IGNITE-12968
> URL: https://issues.apache.org/jira/browse/IGNITE-12968
> Project: Ignite
> Issue Type: Task
> Components: documentation
> Reporter: Maxim Muzafarov
> Assignee: Maxim Muzafarov
> Priority: Blocker
> Labels: iep-43, important
> Fix For: 2.9
>
> Time Spent: 6h 40m
> Remaining Estimate: 0h
>
> Add the following to the Apache Ignite documentation:
> 1. How to create a cluster snapshot (describe API, limitations)
> 2. How to configure a destination directory
> 3. Manual steps for a snapshot restore
> 4. Examples
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
