[
https://issues.apache.org/jira/browse/CASSANDRA-12236?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15404155#comment-15404155
]
Sylvain Lebresne commented on CASSANDRA-12236:
----------------------------------------------
Pushed a commit with fix for those and re-started the tests. I did got a run of
upgrade test to finish previously and there was still 72 failures. I start
looking at the reports and it's not immediately clear than any of those is due
to this issue, but at the same time the reports aren't full of details in many
cases (if we have a node logs, I'm not aware of it). It could be worth noting
there is an overwhelming number of paging tests fails, and it doesn't sound
like paging tests would more sensible to this than other (I've hinted at some
possible explanation for paging upgrade test failures on CASSANDRA-10848 but
I'm not even sure of that), but I'm fishing a bit here.
In any case, I think we want to do this change as it was never intended to ship
columns for which we had no values, but this is probably not the end of our
upgrade test problems.
bq. One nit: the change in MigrationManager is contaminating git blame, no need
for it
Not sure to follow that one. Even if the change wasn't pulling the
{{Collections.singletonList()}} call out, which I happen to think is slightly
cleaner (and thus prefer over {{git blame}} concerns), the code would still
need to call the {{build()}} method so the pollution would be the same. I'm
surely missing your point though.
> RTE from new CDC column breaks in flight queries.
> -------------------------------------------------
>
> Key: CASSANDRA-12236
> URL: https://issues.apache.org/jira/browse/CASSANDRA-12236
> Project: Cassandra
> Issue Type: Bug
> Reporter: Jeremiah Jordan
> Assignee: Sylvain Lebresne
> Fix For: 3.8, 3.9
>
> Attachments: 12236.txt
>
>
> This RTE is not harmless. It will cause the internode connection to break
> which will cause all in flight requests between these nodes to die/timeout.
> {noformat}
> - Due to changes in schema migration handling and the storage format
> after 3.0, you will
> see error messages such as:
> "java.lang.RuntimeException: Unknown column cdc during
> deserialization"
> in your system logs on a mixed-version cluster during upgrades. This
> error message
> is harmless and due to the 3.8 nodes having cdc added to their schema
> tables while
> the <3.8 nodes do not. This message should cease once all nodes are
> upgraded to 3.8.
> As always, refrain from schema changes during cluster upgrades.
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
