Mailing-List: contact commits-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@cassandra.apache.org
Date: Thu, 4 Jun 2015 19:07:39 +0000 (UTC)
From: "Joshua McKenzie (JIRA)" <jira@apache.org>
To: commits@cassandra.apache.org
Message-ID: <JIRA.12830796.1431976284000.14683.1433444859309@Atlassian.JIRA>
In-Reply-To: <JIRA.12830796.1431976284000@Atlassian.JIRA>
References: <JIRA.12830796.1431976284000@Atlassian.JIRA>
 <JIRA.12830796.1431976284095@arcas>
Subject: [jira] [Updated] (CASSANDRA-9418) Fix dtests on Windows
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/jira/browse/CASSANDRA-9418?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Joshua McKenzie updated CASSANDRA-9418:
---------------------------------------
    Attachment: 9418_tz_formatting.txt

Issues w/json tests are two-fold:
# <BLANKLINE> comparisons are mismatched on Windows
# Default timestamp formatting is different on Windows.

Re: #1: python coerces line-endings on all print or write calls to platform-specific endings, CRLF in Windows' case, which subsequently causes failures during the regex change and comparison in doctest.check_output as \s$ doesn't match the \s^M$ on Windows output. In order to resolve this, I've opened a dtest PR to convert CRLF to LF in cqlsh_print on Windows: [PR|https://github.com/riptano/cassandra-dtest/pull/312]

Re: #2: It turns out there's a bug in the interaction between python's strftime and Windows clibs, specifically w/regards to %z in strftime ([bug report|http://bugs.python.org/issue20010]) giving us a long string return on %z rather than the desired +/- vs. GMT.  CASSANDRA-4746 dealt with a somewhat similar issue so I've piggybacked on that and attached a patch that will change the datetime formatter in cqlshlib to manually fix the TZ when on Windows and with a %z formatting flag.

With the PR for dtest and the patch attached, all but 2 of the json tests are passing locally now.

> Fix dtests on Windows
> ---------------------
>
>                 Key: CASSANDRA-9418
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9418
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Joshua McKenzie
>            Assignee: Joshua McKenzie
>              Labels: Windows
>             Fix For: 2.2.x
>
>         Attachments: 9418_tz_formatting.txt
>
>
> There's a variety of infrastructural failures within dtest w/regards to windows that are causing tests to fail and those failures to cascade.
> Error: failure to delete commit log after a test / ccm cluster is stopped:
> {noformat}
> Traceback (most recent call last):
>   File "C:\src\cassandra-dtest\dtest.py", line 452, in tearDown
>     self._cleanup_cluster()
>   File "C:\src\cassandra-dtest\dtest.py", line 172, in _cleanup_cluster
>     self.cluster.remove()
>   File "build\bdist.win-amd64\egg\ccmlib\cluster.py", line 212, in remove
>     shutil.rmtree(self.get_path())
>   File "C:\Python27\lib\shutil.py", line 247, in rmtree
>     rmtree(fullname, ignore_errors, onerror)
>   File "C:\Python27\lib\shutil.py", line 247, in rmtree
>     rmtree(fullname, ignore_errors, onerror)
>   File "C:\Python27\lib\shutil.py", line 252, in rmtree
>     onerror(os.remove, fullname, sys.exc_info())
>   File "C:\Python27\lib\shutil.py", line 250, in rmtree
>     os.remove(fullname)
> WindowsError: [Error 5] Access is denied: 'c:\\temp\\dtest-4rxq2i\\test\\node1\\commitlogs\\CommitLog-5-1431969131917.log'
> {noformat}
> Cascading error: implication is that tests aren't shutting down correctly and subsequent tests cannot start:
> {noformat}
> 06:00:20 ERROR: test_incr_decr_super_remove (thrift_tests.TestMutations)
> 06:00:20 ----------------------------------------------------------------------
> 06:00:20 Traceback (most recent call last):
> 06:00:20   File "D:\jenkins\workspace\trunk_dtest_win32\cassandra-dtest\thrift_tests.py", line 55, in setUp
> 06:00:20     cluster.start()
> 06:00:20   File "build\bdist.win-amd64\egg\ccmlib\cluster.py", line 249, in start
> 06:00:20     p = node.start(update_pid=False, jvm_args=jvm_args, profile_options=profile_options)
> 06:00:20   File "build\bdist.win-amd64\egg\ccmlib\node.py", line 457, in start
> 06:00:20     common.check_socket_available(itf)
> 06:00:20   File "build\bdist.win-amd64\egg\ccmlib\common.py", line 341, in check_socket_available
> 06:00:20     raise UnavailableSocketError("Inet address %s:%s is not available: %s" % (addr, port, msg))
> 06:00:20 UnavailableSocketError: Inet address 127.0.0.1:9042 is not available: [Errno 10013] An attempt was made to access a socket in a way forbidden by its access permissions
> 06:00:20 -------------------- >> begin captured logging << --------------------
> 06:00:20 dtest: DEBUG: removing ccm cluster test at: d:\temp\dtest-a5iny5
> 06:00:20 dtest: DEBUG: cluster ccm directory: d:\temp\dtest-dalzcy
> 06:00:20 --------------------- >> end captured logging << ---------------------
> {noformat}
> I've also seen (and am debugging) an error where a node just fails to start via ccm.
> I'll update this ticket with PR's to dtest or other observations of interest.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)