OpenStack Object Storage (swift)

  1. Bug #1699114
  2. Comment #2

Comment 2 for bug 1699114

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to swift (master)

Reviewed: https://review.openstack.org/507964
Committed: https://git.openstack.org/cgit/openstack/swift/commit/?id=c6aea4b3730c937c41815831a7b4d60ff2899fcb
Submitter: Jenkins
Branch: master

commit c6aea4b3730c937c41815831a7b4d60ff2899fcb
Author: Tim Burke <email address hidden>
Date: Wed Sep 27 19:19:53 2017 +0000

    Fix intermittent failure in test_x_delete_after

    X-Delete-After: 1 is known to be flakey; use 2 instead.

    When the proxy receives an X-Delete-After header, it automatically
    converts it to an X-Delete-At header based on the current time. So far,
    so good. But in normalize_delete_at_timestamp we convert our

        time.time() + int(req.headers['X-Delete-After'])

    to a string representation of an integer and in the process always round
    *down*. As a result, we lose up to a second worth of object validity,
    meaning the object server can (rarely) respond 400, complaining that the
    X-Delete-At is in the past.

    Change-Id: Ib5e5a48f5cbed0eade8ba3bca96b26c82a9f9d84
    Related-Change: I643be9af8f054f33897dd74071027a739eaa2c5c
    Related-Change: I10d3b9fcbefff3c415a92fa284a1ea1eda458581
    Related-Change: Ifdb1920e5266aaa278baa0759fc0bfaa1aff2d0d
    Related-Bug: #1597520
    Closes-Bug: #1699114