openstack/swift - swift - OpenDev: Free Software Needs Free Tools

* ensure all responses are 200 OK
* report missing sub-SLO manifests or other error messages in bulk
 delete response
Change-Id: Iaf88c94bc7114ff3c9751f9f31f8f748de911f8a

assertTrue accepts a parameter msg which will be printed when
assertion fails, usually msg is a str. This patch fixes unsuitable
usage of assertTrue which set msg to bool type True.
Change-Id: I731f8ea553c935eba0e112ffded16f41a5ea86c0
Fixes-Bug: #1226374 

Change-Id: I21175a4be0cda9a8a98c425bff11c80895cd6d3e

Change-Id: Ia28a9b47213f848ab5ea59572e14ac710ed881e3

Change-Id: I33d9119684497b53101db3bc380953e86bdf25f0

Change-Id: I03573bf24baf031b1874c3ff2e2d89d34473c266

Change-Id: I7c3df47c31759dbeb3105f8883e2688ada848d58
Closes-bug: #1214176 

RFC 1945 and RFC 2068 specify that the client is not allowed to change the
method on the redirected request. However, most existing user agent
implementations treat 302 as if it were a 303 response, performing a GET on
the Location field-value regardless of the original request method. The status
codes 303 and 307 have been added for servers that wish to make unambiguously
clear which kind of reaction is expected of the client.
HTTP/1.1 RFC for 302: http://tools.ietf.org/html/rfc2616#section-10.3.4
HTTP/1.1 RFC for 307: http://tools.ietf.org/html/rfc2616#section-10.3.8
Change-Id: I354e2f4f3e3eb6a1553b3d9c60b613d8f0c37531
Signed-off-by: Prashanth Pai <ppai@redhat.com>

Place all the methods related to on-disk layout and / or configuration
into a new common module that can be shared by the various modules
using the same on-disk layout.
Change-Id: I27ffd4665d5115ffdde649c48a4d18e12017e6a9
Signed-off-by: Peter Portante <peter.portante@redhat.com>

Change-Id: I957deeb0e711bfe7cd9d852726c77179a4613ee0

This is a common configuration option and allows devstack (and others)
to configure a mod_wsgi user to run the daemon process.
Change-Id: Idf134b3bc6b08e3c3a80dde8830d5a4f3da5a06c
Fixes: bug 1226346

Hopefuly this extra test case is not too inane and slows us
down for nothing. It is verified to fail with the old code.
Change-Id: I604eca09f7f9ae044a8ad75284cd82a37325f99c

This has a few advantages: some of these locations will perform
slightly faster, it's minutely less code, and dict.keys has a different
behavior on Python 3. Also removes usage of dict.iterkeys, which is
identical to just iterating over the dict, except it doesn't exist on
Python 3 at all.
Change-Id: Ia91e97232dc8d78cf63fa807de288fc25cf5425d

When we're looking for handoffs, we first try to span all the regions,
then zones, then devices. However, the search loops would always go
over a big subset of the ring, regardless of whether or not there was
anything left to find.
In particular, this would result in long runtimes for the first call
to get_more_nodes() if you didn't have more regions than replicas, as
the region-search loop would chew through a ton of devices and not
find any (since there weren't any). Now we do a little arithmetic,
think about the pigeonhole principle a little bit, and bail out when
there's nothing left to find, rather than waiting until there's no
space left to search.
Similar changes were made for the different-zone and different-device
search loops.
On a 4800-device test ring (4 regions, 5 zones each, 20 node each, 12
drives each), the time to get all handoffs dropped to about 5% of its
previous value (a 95% improvement).
More usefully, on a 1200-device test ring (same as above but only 1
region), the time to get just the first 6 handoffs dropped to about
0.01% of its original runtime, or a 10,000x speedup.
Fixes bug 1225018.
Change-Id: I4c77094186f0032a3e19a099a1a0e71b2ba06719

This patch is only a cleanup, but our code makes me rage every time
I read it. In short, we have a class with a variable self.auditor,
and we assign ObjectAuditor to it at some ties, AuditorWorker at
other times. So, whenever there's a mismerge or whatever, you cannot
tell if self.auditor.broker_class makes sense or not.
Since all cases of using self.auditor as AuditorWorker are purely
local, just evict it to a local variable auditor_worker. Leave
self.auditor to represent ObjectAuditor, because it's used as
a class variable in places.
Change-Id: I811b80ac6c69a4daacfed7a5918bc0b15761d72e

Implicit relative imports don't work on Python3, and are also prone to
several different classes of errors.
Change-Id: I7b62e9bfbe9c0b1fc9876413e3139fda019a4e57

This should make ``Ring.get_more_nodes`` microscopically faster.
Change-Id: Ibf0988fe0630ad94ac0c04040766d89ef86d1488

If handoffs_first is True, then the object replicator will give
partitions that are not supposed to be on the node priority.
If handoff_delete is set to a number (n), then it will delete a handoff
partition if at least n replicas were successfully replicated
Also fixed a couple of things in the object replicator unit tests and
added some more
DocImpact
Change-Id: Icb9968953cf467be2a52046fb16f4b84eb5604e4

Change-Id: I69a14bcb92490abd3ad1070799bf1580a2dcaa92
Signed-off-by: Peter Portante <peter.portante@redhat.com>

Because the audit code catches all Exceptions, bugs in the code, like
misspelled names, bad logic, can result in errors which don't cause
exceptions that kill the unit tests.
Change-Id: Idb8e3b00a49863ab920161a46edb61a5c47fffe2

Change-Id: Id232c6973dcb55c19233b6d517770e58c2737ba0

fix bug 1224208
Change-Id: I83c673a87c31214a7c54b6399ca53512885e6bc3

Change-Id: If80509e2e19cec5d0b8c5cfccd15e6d893558de4

- When localhost:80 was binding the tests was trying to connect into it.
- To test you can simply run sudo python -m SimpleHTTPServer 80 which
 should show :
 1.0.0.127.in-addr.arpa - - [06/Aug/2013 14:10:42] code 501, message Unsupported method ('DELETE')
 1.0.0.127.in-addr.arpa - - [06/Aug/2013 14:10:42] "DELETE /a/c/o HTTP/1.1" 501 -
 (the test was passing since 501 would raise ClientException).
 mock delete_object in the fourth test to fix that
- Refactor the code to use mock.patch as well.
Closes-Bug: 1208802
Change-Id: I5ddd4ac3a97879f51cf5883fcfc0fe0f0adaeff6

* test/__init__.py: Put safe_repr import/implementation here so that it
is available to functional and unit tests.
* test/functional/swift_test_client.py: When a request fails
record why that request failed, how many requests failed, and what the
request was when raising RequestError to aid in debugging. Makes use of
safe_repr from test/__init__.py.
* test/unit/common/test_constraints.py: Remove implementation of
safe_repr and use the implementation in test/__init__.py.
Change-Id: I6c957343fb4b8b95d3875fd5ca87b3cf28a5f47a

Used groff to recreate the errors. I believe all the issues
except `binary-without-manpage` are solved. Would like
confirmation from someone using Lintian.
Closes-Bug: #1210114
Change-Id: I533205c53efdb7cdf3645cc3e3dc487f9ee5640a

Add more unittests for account_reaper in order to
improve code coverage.
Fixes: bug #1210045
Change-Id: I73490f80ac70b814a0abc490dc932f2a86fca703