[
https://issues.apache.org/jira/browse/IGNITE-15492?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17429181#comment-17429181
]
Alexander Lapin commented on IGNITE-15492:
------------------------------------------
[~v.pyatkov] I've checked the PR, first of all, as we've discussed, let's
validate whether we should return consistent data from node B in case of
altering table from node A. I believe that we should.
In addition to that assuming that given solution is a temporary one, seems that
proposed variant will be extremely in-effective cause we we'll check MS on
every get. Vlad did you consider other options that'll be more effective?
It also worth to actualize code. For example
[IGNITE-15412|https://issues.apache.org/jira/browse/IGNITE-15412] is already
implemented, so it worth to rewrite _latestSchemaVersion()_
> Check schema availability on local node
> ---------------------------------------
>
> Key: IGNITE-15492
> URL: https://issues.apache.org/jira/browse/IGNITE-15492
> Project: Ignite
> Issue Type: Sub-task
> Reporter: Vyacheslav Koptilin
> Assignee: Vladislav Pyatkov
> Priority: Major
> Labels: ignite-3
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> Currently, we do not check that a schema that is needed for deserialization
> returned tuple (for example, the end-user calls tableKvView.get(key) method)
> is available on the local node that performs table operation.
> Need to implement a synchronization mechanism that allows overcoming an issue
> related to the deserialization of tuples. Looks like it is an actual issue
> regardless of rebalancing.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
