Github user iamaleksey commented on the issue:
https://github.com/apache/cassandra/pull/239
@dineshjoshi nice round of changes! I think we are there, or almost-almost
there, at worst. Let me do one last pass before we go ahead and commit.
---
Github user dineshjoshi commented on the issue:
https://github.com/apache/cassandra/pull/239
@iamaleksey made a few more changes -
1. Got rid of `IStreamWriter`
2. Ensured we're logging the configuration warning only once at start up
iff zero copy streaming is enabled
Github user dineshjoshi commented on the issue:
https://github.com/apache/cassandra/pull/239
@iamaleksey I've addressed your comments including the one about disabling
faster streaming for legacy counter shards.
I did add a much less expensive check for STCS. It won't get all
Github user dineshjoshi commented on the issue:
https://github.com/apache/cassandra/pull/239
@iamaleksey I *think* I've resolved all your comments. The last dtest was
all green. I'm not sure about the latest push but I don't forsee any breakages.
Please let me know if there are more
Github user dineshjoshi commented on the issue:
https://github.com/apache/cassandra/pull/239
@iamaleksey I have resolved most of your comments. I still have a couple to
go. I will update this PR when I am done with those and get a clean dtest run.
---
Github user aweisberg commented on the issue:
https://github.com/apache/cassandra/pull/239
Where is the dtest? The existing link doesn't seem to work and when I
looked at your repo there are no branches with these tests?
---
Github user dineshjoshi commented on the issue:
https://github.com/apache/cassandra/pull/239
@aweisberg I have updated the dtest to test with different compaction
strategies and addressed remaining comments.
---
-
Github user aweisberg commented on the issue:
https://github.com/apache/cassandra/pull/239
Ah it's the log statement block stream? Great. I think the thing I would
look for then is just some log details about what percentage was block streamed
vs regular streamed and assert that the
Github user dineshjoshi commented on the issue:
https://github.com/apache/cassandra/pull/239
1. Fixed the data directory selection
2. Added & updated unit tests
3. Added documentation in cassandra.yaml
4. Updated the microbenchmark to benchmark the read and write path for
9 matches
Mail list logo