commit c004a09c7ddd596a4f608e7152fe84988b0ebd3e
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Wed Mar 1 02:09:40 2017 +1000
Automatic commit of release 24.1
commit 5f1e023ead31c484cc88e740a0f82bc8fff97438
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Tue Feb 28 21:30:04 2017 +1000
release notes for 24.1
Change-Id: I1bc911b9d5f3bf33c10dc1510eb4d0d2271c9929
commit 9d75df3967b690cf96987949ef58f60f940d6b40
Author: Matt Jia <mjia@redhat.com>
Date: Thu Feb 2 16:07:17 2017 +1000
update the mapping of Beaker version to db schema version for 24.0
Change-Id: I436272b7500a55aef7a5187172060aef313cf76f
commit 80371d48050e0e8333141a2b90e3a77ced9c75ab
Author: Matt Jia <mjia@redhat.com>
Date: Thu Feb 2 15:46:57 2017 +1000
remove the special handling for the data migrations
Since we are in 24.0, we do not need to handle the case in which the
data_migration table doesn't exist
Change-Id: I03c3bcba091e6d02ee09b0c0694ca27fd293cefd
commit 04ccf7abc458f897d1adf988647cca4f760fb8d6
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Sat Feb 25 07:29:50 2017 +1000
fix a dumb typo in commit 10b69af8
Change-Id: Ic1429510eaf87c5e8e94889e418924c740dac238
commit bd17690b324bf01aac695aed0d1fc9f6a8b335e4
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Wed Feb 22 07:38:15 2017 +1000
only show link to http://$THESYSTEM if recipe is Running or Reserved
The purpose of the link is to be a shortcut to some web application
which is set up as part of the recipe, so it only makes sense to show
the link once the installation is done, and while the recipe is not
finished.
Bug: 1412878
Change-Id: Ib9b42e7ae57f36201cb2c85d107c80ec4d5af4e2
commit 10b69af8bc22d2345d28d818b2a9410a94090488
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Wed Feb 22 02:53:07 2017 +1000
send beakerd metrics queries to reports engine
These queries are somewhat expensive, and they don't need to be
totally live, so they are a good candidate for offloading to a separate
replica database.
Bug: 1425522
Change-Id: Idd329977b67c07209777c4604b5b5d5441ce71a5
commit 981a6c2df1eb9383311a7f3b3354dd7ac6f53bc6
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Tue Feb 21 02:05:15 2017 +1000
fix typo in populating excluded_osmajors for task uploads
Bug: 1422410
Change-Id: I3409126ceac67d8f74340316f31f56347649e67c
commit 80793a3e6649593ee3f4855c0eec5c04df4fbe2a
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Tue Feb 21 02:00:16 2017 +1000
tests: ensure we start with a pristine tasks directory
When the test suite starts the database is empty (there are no task
records) so let's make the basepath.rpms directory match.
Currently, if there are some files left behind in the basepath.rpms
directory from a previous test run, it will cause some tests in
bkr.inttest.client.test_task_add to fail.
The tests in bkr.inttest.server.selenium.test_tasks were avoiding this
problem by unlinking the files as soon as they were uploaded, but that
is actually backwards. A more realistic scenario is that the files are
*not* on disk for a fresh Beaker with an empty database, and they *are*
left on disk after uploading.
Change-Id: I93490a46c23e3e9e46bfa2c8252448d331a47675
commit 83d9e961627a285d8b4a36a297b6cced1c48f69d
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Fri Feb 17 00:49:45 2017 +1000
also log *before* handling an XMLRPC request
Currently the only way we can see what was in an XMLRPC request is the
debug logging which Beaker produces. The Apache access logs are useless
for XMLRPC because every request is just POST /client/.
However if the request hits a MemoryError this typically breaks all the
error handling code paths as well, which means that the log message at
the end is never logged and so it's impossible to know which XMLRPC
request triggered the MemoryError.
If we also log the details of the XMLRPC request *before* we start
handling it we should be able to identify the requests even if they
subsequently hit MemoryError.
Change-Id: Ib57c4f5a217ce4b21212214ec4fe985088571b76
commit 536e9acc05cd32d6e36180104283b446b3976d02
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Fri Feb 17 01:24:22 2017 +1000
avoid allocations in the last resort error handling path
This error handling path is supposed to kill Apache workers that have
hit MemoryError before they poison any other requests. But just logging
a message can itself crash with MemoryError.
Bug: 1422874
Change-Id: I3072c4d334d01b7396ce988f34eddcec3f1a4f72
commit 3674d26266a32f75790c943fee43f29b09695b36
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Wed Jan 18 12:36:27 2017 +1000
ignore 'src' arch when importing Fedora 25
Bug: 1412487
Change-Id: Ifcb13fd7d9398a502309c4459330d37a13f4d50b
commit faf83f4aa9cbc18af1aebf6df5d9710921b17194
Author: Dan Callaghan <dcallagh@redhat.com>
Date: Thu Feb 9 19:21:00 2017 +1000
docs: fix wrong indentation in 24.0 release notes
The extra space was being interpreted as a blockquote in the middle of
the paragraph.
Change-Id: Ifa8908d03702b1ee1d08d323e405acc08b58128a