cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Stefania (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-11701) [windows] dtest failure in cqlsh_tests.cqlsh_copy_tests.CqlshCopyTest.test_reading_with_skip_and_max_rows
Date Fri, 05 Aug 2016 06:52:20 GMT


Stefania commented on CASSANDRA-11701:

I rewrote the patch to avoid using a thread lock, since it cannot be pickl-ed on Windows.
I extracted the pipe out of the channels, so we only transfer pipes and not channels to the
child processes, which then create the channels after forking. To avoid redundant threads,
which are only required for sending, I've separated receiving from sending channels. Only
sending channels create a feeding thread, right in the constructor so the lock is no longer

I'm not sure if we want this fix in 2.1 or not, it is not a critical bug but it is a regression
compared to the old cqlsh COPY functionality. It is a rare failure but it can occur if the
main thread of a child process needs to send an error when the receiving thread is already
sending results.


The windows test results will be available [here|],
build #7 and #8.

> [windows] dtest failure in cqlsh_tests.cqlsh_copy_tests.CqlshCopyTest.test_reading_with_skip_and_max_rows
> ---------------------------------------------------------------------------------------------------------
>                 Key: CASSANDRA-11701
>                 URL:
>             Project: Cassandra
>          Issue Type: Test
>            Reporter: Russ Hatch
>            Assignee: Stefania
>              Labels: dtest, windows
> looks to be an assertion problem, so could be test or cassandra related:
> e.g.:
> {noformat}
> 10000 != 331
> {noformat}
> Failed on CassCI build trunk_dtest_win32 #404

This message was sent by Atlassian JIRA

View raw message