[
https://issues.apache.org/jira/browse/CASSANDRA-15694?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17089293#comment-17089293
]
Dinesh Joshi commented on CASSANDRA-15694:
------------------------------------------
Hi [~stefan.miklosovic], thanks for the bug report and patch. Here's my
feedback:
# {{StreamSession::addTransferStreams}} doesn't seem to be the right place to
compute the number of files inside an {{OutgoingStreamMessage}}. It belongs to
the {{OutgoingStreamMessage}}.
# While you're correctly sending the number of files in the stream message on
the outbound side, on the receiving side you did not update the number of files
that the receiver expects. This causes mismatch and the receiver keeps waiting
to receive files specifically {{StreamReceiveTask}} hangs.
# A few existing tests broke as the number of files they expected changed,
specifically tests in {{StreamTransferTaskTest}}.
# Minor spacing nits in test file.
I ran CircleCI and spotted a few tests failed due to timeouts and thats how I
was able to figure out that things didn't function as expected.
I have made all suggestions in this
[branch|https://github.com/dineshjoshi/cassandra/tree/CASSANDRA-15694-trunk-review].
Please let me know your thoughts. If it looks good, I'll go ahead and commit
it.
> Statistics upon streaming of entire SSTables in Netstats is wrong
> -----------------------------------------------------------------
>
> Key: CASSANDRA-15694
> URL: https://issues.apache.org/jira/browse/CASSANDRA-15694
> Project: Cassandra
> Issue Type: Bug
> Components: Tool/nodetool
> Reporter: Stefan Miklosovic
> Assignee: Stefan Miklosovic
> Priority: Normal
> Time Spent: 10m
> Remaining Estimate: 0h
>
> There is a bug in the current code (trunk on 6th April 2020) as if we are
> streaming entire SSTables via CassandraEntireSSTableStreamWriter and
> CassandraOutgoingFile respectively, there is not any update on particular
> components of a SSTable as it counts only "db" file to be there. That
> introduces this bug:
>
> {code:java}
> Mode: NORMAL
> Rebuild 2c0b43f0-735d-11ea-9346-fb0ffe238736
> /127.0.0.2 Sending 19 files, 27664559 bytes total. Already sent 133
> files, 27664559 bytes total
>
> /tmp/dtests15682026295742741219/node2/data/distributed_test_keyspace/cf-196b3...
> ....
> {code}
> Basically, number of files to be sent is lower than files sent.
>
> The straightforward fix here is to distinguish when we are streaming entire
> sstables and in that case include all manifest files into computation.
>
> This issue depends on CASSANDRA-15657 because the resolution whether we
> stream entirely or not is got from a method which is performance sensitive
> and computed every time. Once CASSANDRA-15657 (hence CASSANDRA-14586) is
> done, this ticket can be worked on.
>
> branch with fix is here:
> [https://github.com/smiklosovic/cassandra/tree/CASSANDRA-15694]
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
