Reviewed:  https://review.openstack.org/530733
Committed: https://git.openstack.org/cgit/openstack/swift/commit/?id=e2f780492487a31d4f8111d53cbde6c02d9d3237
Submitter: Zuul
Branch:    feature/deep

commit 61fe6aae81d00597c777a64ac337a8dfb990f0c2
Author: Tim Burke <email address hidden>
Date:   Tue Aug 22 22:40:58 2017 +0000

    Better mock out OSErrors in test_replicator before raising them
    
    Also, provide a return value for resp.read() so we hit a
    pickle error instead of a type error.
    
    Change-Id: I56141eee63ad1ceb2edf807432fa2516fabb15a6

commit 0bdec4661b5609ca1bf813a7ccd514e5d444b07f
Author: Kazuhiro MIYAHARA <email address hidden>
Date:   Mon Dec 25 09:13:17 2017 +0000

    Skip symlink + vw functional tests if symlink is not enabled
    
    Functional tests for symlink and versioned writes run and result in
    falure even if symlink is not enabled.
    
    This patch fixes the functional tests to run only if both of
    symlink and versioned writes are enabled.
    
    Change-Id: I5ffd0b6436e56a805784baf5ceb722effdf74884

commit 17e6950aa08101b5f3bec0f2f9c32cfd5f51fa36
Author: Kazuhiro MIYAHARA <email address hidden>
Date:   Fri Dec 22 02:18:09 2017 +0000

    Fix manpage docs' daemon names
    
    In current manpage docs, some of daemon names for concurrency
    explanation is wrong.
    
    This patch fixes the daemon names.
    
    Change-Id: I2a505c9590ee3a3a7e37e8d949a10db36206faec

commit af2c2a6eb54d848eefc2d0a1b619e0b86eed2eb5
Author: Samuel Merritt <email address hidden>
Date:   Thu Dec 21 10:43:39 2017 -0800

    Fix sometimes-flaky container name functional test.
    
    You've got two test classes: TestContainer and TestContainerUTF8. They
    each try to create the same set of containers with names of varying
    lengths to make sure the container-name length limit is being honored.
    
    Also, each test class tries to clean up pre-existing data in its
    setUpClass method. If TestContainerUTF8 fails to delete a contaienr
    that TestContainer made, then its testContainerNameLimit method will
    fail because the container PUT response has status 202 instead of 201,
    which is because the container still existed from the prior test.
    
    I've made the test consider both 201 and 202 as success. For purposes
    of testing the maximum container name length, any 2xx is fine.
    
    Change-Id: I7b343a8ed0d12537659c051ddf29226cefa78a8f

commit 609c757e698ff7893e1b1a0e32d088ad9d05ad95
Author: Clay Gerrard <email address hidden>
Date:   Tue Dec 12 21:39:54 2017 -0800

    functest for symlink + versioned writes
    
    Co-Author: Alistair Coles <email address hidden>
    
    Related-Change-Id: I838ed71bacb3e33916db8dd42c7880d5bb9f8e18
    Change-Id: I0ccff1eafcfb3fdbdda9faf55a44c45b834e723a

commit bdd4eb6936b0e25aff5357bde876309ee5b032ec
Author: Andreas Jaeger <email address hidden>
Date:   Wed Dec 20 07:14:03 2017 +0100

    Install liberasurecode-devel for CentOS 7
    
    Since I747c2b8754effbc6ec82af3bf7543fd9599a6c14 we do not install
    the RDO package repository anymore and thus liberasurecode-devel
    cannot be installed.
    
    For CentOS 7, remove liberasurecode-devel from bindep.txt and install it
    from test-setup.sh instead after enabling the RDO package repositories.
    
    Update python dependencies: CentOS 7 does not have python3. Fix the
    SUSE tags.
    
    Change-Id: I72aa6b5455dfb025f54e83334983ac280f04afb2

commit dc1c55c9a07c03fe85f4bcc52419a42d75ae30fa
Author: Andreas Jaeger <email address hidden>
Date:   Sun Dec 17 19:46:52 2017 +0100

    Native Zuul v3 tox jobs
    
    Convert the legacy tox jobs to Zuul v3 native and use the
    tools/test_setup.sh script to setup a XFS file like it's done in the
    legacy job.
    
    Needed-By: Id2b5cff998ac3a825a8f515c7bae3b433f30d272
    Change-Id: I34ed9e1c4b822f700e42fb07937df7be72cbaf4e

commit a7da2232629bfd4d5c04f5169e51b1f57b6c9362
Author: Matthew Oliver <email address hidden>
Date:   Tue Dec 19 05:47:20 2017 +0000

    Fix intermittent problem in part_swapping test
    
    There is an intermittent failure in the test_part_swapping_problem
    test found in test/unit/common/ring/test_builder.py.
    
    The test does a rebalance, then changes the ring to test a specific
    problem, does some housekeeping and then rebalances again. The problem
    is the ringbuilder keeps track of where in the ring it started the
    last ring rebalance, saved in `_last_part_gather_start`.
    
    On a rebalance, or more specifically in `_gather_parts_for_balance` we
    then we will start somewhere on the other side of the ring with:
    
      quarter_turn = (self.parts // 4)
      random_half = random.randint(0, self.parts / 2)
      start = (self._last_part_gather_start + quarter_turn +
                     random_half) % self.parts
    
    Because we don't reset `_last_part_gather_start` when we change the ring
    in the test, there is edge case where if we are unlucky during both
    rebalances whereby both calls to randint returns a relatively large
    number pushes the start of the second rebalance to the wrong side
    of the ring. Actually it's more problematic, and only 1 large random
    and a one in the middle will cause it, maybe pictures help:
    
      rabalance 1 (r1): quarter_turn = 4, random_half = 5
      rebalance 2 (r2): quarter_turn = 4, random_half = 3
    
                                         r1                   r2
                                          |                    |
                                          v                    v
      array('H', [0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1]),
      array('H', [1, 1, 1, 1, 2, 2, 2, 3, 3, 3, 2, 2, 2, 3, 3, 3]),
      array('H', [2, 2, 2, 2, 3, 3, 4, 4, 4, 4, 3, 4, 4, 4, 4, 4])]
    
    Now when gathering for rebalance 2 it'll pick:
    
      array('H', [0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, X]),
      array('H', [X, X, 1, 1, 2, 2, 2, 3, 3, 3, 2, 2, 2, 3, 3, 3]),
      array('H', [2, 2, 2, 2, 3, 3, 4, 4, 4, 4, 3, 4, 4, 4, 4, 4])]
    
    Which can cause the 3 attempts to gather and rebalance to be used up.
    This causes the intermittent failure seen in the bug.
    
    This patch solves this by resetting `_gather_parts_for_balance` to 0
    while we tidy up the ring change. Meaning we'll always start on the
    correct side of the ring.
    
    Change-Id: I0d3a69620d4734091dfa516efd0d6b2ed87e196b
    Closes-Bug: #1724356

commit 2cf5e7ceffba007d1ff2a429385c1f3994a59d65
Author: John Dickinson <email address hidden>
Date:   Mon Dec 18 09:33:40 2017 -0800

    fix SkipTest imports in functests so they can be run directly by nose
    
    Change-Id: I7ecc48f69ca677d5ecb0986ac4042688442355bb

commit aa82d2cba82209f1bf3944c6d2a67965af5a1540
Author: Samuel Merritt <email address hidden>
Date:   Thu Jun 29 10:23:38 2017 -0700

    Save ring builder if dispersion changes
    
    There are cases where a rebalance improves dispersion, but doesn't
    improve balance. This is because the balance of a ring builder is
    taken to be the balance of its least-balanced device, so if there's a
    device that has no partitions, wants some, but can't get them, then
    we'll never save the ring builder even if every other device in the
    ring got better.
    
    We can detect this situation by looking at the dispersion number; if it
    changes, then the rebalance needs to be saved in order to continue to
    make progress.
    
    Partial-Bug: #1697543
    
    Change-Id: Ie239b958fc7e0547ffda2bebf61546bd4ef3d829

commit 1984353f0d6db7512e4ea147ecad9e14dfb318d4
Author: Alistair Coles <email address hidden>
Date:   Fri Dec 15 12:36:47 2017 +0000

    Move symlink versioning functional test
    
    The functional test for versioning symlinks is better located in
    test_versioned_writes where it can be added to
    TestObjectVersioning. This saves duplicated versioned_writes specific
    setup code in test_symlink, and has the benefit of the test being
    repeated for each of the versioned writes test subclasses.  With a
    small refactor this includes the test now running with
    x-history-location mode as well as x-versions-location mode.
    
    Related-Change: I838ed71bacb3e33916db8dd42c7880d5bb9f8e18
    Change-Id: If215446c558b61c1a8aea37ce6be8fcb5a9ea2f4

commit c579e99126b61466fb3b1628170cbca37ccacce3
Author: Kota Tsuyuzaki <email address hidden>
Date:   Wed Dec 13 06:04:40 2017 +0000

    Add more assertions for Symlink + Copy unit tests
    
    Related-Change-Id: I838ed71bacb3e33916db8dd42c7880d5bb9f8e18
    Change-Id: Ib4c8f0c83537b74bbdec8c2dd6acc99c039faa67

commit 097e975938befb939fa6821f50c061e2c7f42cef
Author: Clay Gerrard <email address hidden>
Date:   Thu Dec 14 16:17:29 2017 -0800

    Remove symlink from xml listing response
    
    We've had some problems with brittle XML clients in the past - it might
    be safer to ask that clients that need symlink keys in listings from
    containers request in JSON.
    
    Change-Id: I4ac7457f3ccb10f9e471ec6dc6f0869d71712878

commit 7647defb0f201cbf85baef6404067bb0eb27321f
Author: Clay Gerrard <email address hidden>
Date:   Thu Dec 14 12:56:49 2017 -0800

    rename utils function less like stdlib
    
    Related-Change-Id: I3436bf3724884fe252c6cb603243c1195f67b701
    Change-Id: I74199c62b46e4db93a76760ebf91d84e3e1e3cfc

commit b342a8147c38ebcf02c3ba21fd09ded4ca49f69b
Author: Alistair Coles <email address hidden>
Date:   Tue Dec 12 17:32:55 2017 +0000

    Assert X-Newest and X-Backend headers are propagated to symlink target
    
    Adds some assertions to verify that X-Newest and X-Backend-* headers are
    copied from an original object GET requets to the symlink target
    request.
    
    Change-Id: Idce92edd002dec34f5dbc5d3c28a4cbbd2fbdc60
    Related-Change: I838ed71bacb3e33916db8dd42c7880d5bb9f8e18

commit 1d5cf3e73067751c8c8fd4f7f58c205db9b877a1
Author: Clay Gerrard <email address hidden>
Date:   Thu Dec 14 12:15:19 2017 -0800

    add symlink to probetest for reconciler
    
    Change-Id: Ib2c5616f2965ab92b1c76d573e869206c91464c6

commit 90134ee968f6f6442eedb5548ee292fc03c77c2a
Author: Clay Gerrard <email address hidden>
Date:   Thu Dec 14 12:09:04 2017 -0800

    add symlink to container sync default and sample config
    
    Change-Id: I83e51daa994b9527eccbb8a88952c630aacd39c3

commit 8df263184be136f9ab203a2971b4f47a52f8b431
Author: Alistair Coles <email address hidden>
Date:   Tue Dec 12 17:26:03 2017 +0000

    Symlink doc clean up
    
    Cleanup for docs and docstrings.
    
    Related-Change: I838ed71bacb3e33916db8dd42c7880d5bb9f8e18
    Change-Id: Ie8de0565dfaca5bd8a5693a75e6ee14ded5b7161

commit 1b6842deafe485bc87c48e1797997d599e7411c9
Author: Mahati Chamarthy <email address hidden>
Date:   Thu Dec 14 14:01:13 2017 +0530

    add name to core emeritus
    
    Change-Id: Icab1c646ec8c9062580197482b1fd924bbc6c4bd

commit 99b89aea1051208e3d71afa35fcd62035e702628
Author: Robert Francis <email address hidden>
Date:   Wed Oct 7 15:14:58 2015 -0400

    Symlink implementation.
    
    Add a symbolic link ("symlink") object support to Swift. This
    object will reference another object. GET and HEAD
    requests for a symlink object will operate on the referenced object.
    DELETE and PUT requests for a symlink object will operate on the
    symlink object, not the referenced object, and will delete or
    overwrite it, respectively.
    POST requests are *not* forwarded to the referenced object and should
    be sent directly. POST requests sent to a symlink object will
    result in a 307 Error.
    
    Historical information on symlink design can be found here:
    https://github.com/openstack/swift-specs/blob/master/specs/in_progress/symlinks.rst.
    https://etherpad.openstack.org/p/swift_symlinks
    
    Co-Authored-By: Thiago da Silva <email address hidden>
    Co-Authored-By: Janie Richling <email address hidden>
    Co-Authored-By: Kazuhiro MIYAHARA <email address hidden>
    Co-Authored-By: Kota Tsuyuzaki <email address hidden>
    
    Change-Id: I838ed71bacb3e33916db8dd42c7880d5bb9f8e18
    Signed-off-by: Thiago da Silva <email address hidden>

commit b86bf15a644db4438770801a312fe074a09c91ef
Author: Thiago da Silva <email address hidden>
Date:   Mon Dec 11 06:59:59 2017 -0500

    remove left-over old doc in config file
    
    Removing the rest of fast post documentation
    from proxy configuration file.
    
    Change-Id: I6108be20606f256a020a8583878e2302e3c98ff1

commit 8182fa75a167d034e29c04680502ea9877592c9a
Author: Samuel Merritt <email address hidden>
Date:   Fri Dec 8 14:08:25 2017 -0800

    Fix small error in a doc string
    
    Change-Id: I1c743fdea637ce047d09a49db0a43b2eb37305fa

commit dd113ab25a3089fa19c8e824c2d89db0ca3db0fa
Author: Alistair Coles <email address hidden>
Date:   Tue Dec 5 10:36:38 2017 -0800

    Refactor proxy-server conf loading to a helper function
    
    There were two middlewares using a common pattern to load
    the proxy-server app config section. The existing pattern
    fails to recognise option overrides that are declared using
    paste-deploy's 'set' notation, as illustrated by the change
    to test_dlo.py in this patch.
    
    This patch replaces the existing code with a helper function
    that loads the proxy-server config using the paste-deploy loader.
    The resulting config dict is therefore exactly the same as that
    used to initialise the proxy-server app.
    
    Change-Id: Ib58ce03e2010f41e7eb11f1a6dc78b0b7f55d466

commit 84ea58b8c81814a3c4d450145bfb9e70166dd60b
Author: Christopher Bartz <email address hidden>
Date:   Mon Dec 4 14:53:44 2017 +0100

    Ringbuilder: Forbid writing empty rings
    
    Swift definitely can't make any use of empty rings, so it should
    not be allowed to write them.
    
    Replace warning with an error message & error exit.
    
    Change-Id: I3a1b86368d363e67d1f91d7d8af4b391a0a53fff
    Closes-Bug: #1396841

commit fc12d63c76ffd51a5d524caa92b78d29bc4e6a7d
Author: Alistair Coles <email address hidden>
Date:   Wed Dec 6 10:29:52 2017 -0800

    Remove repeated text from deployment guide
    
    Fix what appears to be a cut and paste error.
    
    Change-Id: Iccf97ebbf75c8f97095a4493ea6a8beb074df099

commit bc2e03d1a6cb1411a2eee8218cce53f951100eb1
Author: Alistair Coles <email address hidden>
Date:   Tue Dec 5 13:45:18 2017 -0800

    Add --swift-versions option to swift-recon man page
    
    Related-Change: I3c2e569f0c44168333251bb58bab4b5582e15a45
    
    Change-Id: I9776c0919164e48ac445eacf7d897a23ef8e4572

commit 206f674014fb94a63ebdae91e3eca38956055b07
Author: Peter Lisák <email address hidden>
Date:   Tue Dec 20 16:27:36 2016 +0100

    Added swift version to recon cli
    
    Show swift version and check if the same on hosts by using
    `swift-recon --swift-version`.
    
    Ex:
    $ swift-recon --swift-versions
    
    Versions matched (2.7.1.dev144), 0 error[s] while checking hosts.
    
    or if differs
    
    Versions not matched (2.7.1.dev144, 2.7.1.dev145),
    0 error[s] while checking hosts.
    
    Change-Id: I3c2e569f0c44168333251bb58bab4b5582e15a45

commit 924f0d28e9670163dc47206cc1472ef228fd0082
Author: Tim Burke <email address hidden>
Date:   Wed Nov 22 16:51:06 2017 -0800

    dlo: Move conn2 business to the one test that uses it
    
    ...and skip it if we don't have the required user
    
    Change-Id: I8700a587d5b8acff1f0255529b6ddaeadaaa6548

commit 396380f3406bdcd842e85a66cb346bf219cbb4f3
Author: Tim Burke <email address hidden>
Date:   Thu Nov 2 11:43:16 2017 -0700

    Better handle missing files in _construct_from_data_file
    
    There's a window between when we list the files on disk and when we actually
    try to open the .data file where another process could delete it. That should
    raise a DiskFileNotExist error, not an IOError.
    
    Change-Id: I1b43fef35949cb6f71997874e4e6b7646eeec8fe
    Closes-Bug: 1712662

commit 31dd95ee91589be937565692a1610ba49db29706
Author: Tim Burke <email address hidden>
Date:   Thu Oct 12 22:27:35 2017 +0000

    Add base64decode function to common/utils
    
    Keymaster middleware does some nice input validation on
    base64-encoded strings; pull that out somewhere common so
    other things (like SLOs with inlined data) can use it, too.
    
    Change-Id: I3436bf3724884fe252c6cb603243c1195f67b701