ceph: fix the metric list to process with new measures
Currently, the list returned in the Ceph driver contains a lot of
doubloons because it returns a list and not a set. If 1 metric has N new
measures batch waiting to be processed, the list return will be size N
and not 1.
Using a set() avoids that issue and the memory draining implied.
Closes-Bug: #1533793
Change-Id: I3a0b726aae14a17a23a365babc1a2537fb4d1052
(cherry picked from commit 614e13d47fdcaeea9d41bebc214014e0c83a0e83)
Reviewed: https:/ /review. openstack. org/269531 /git.openstack. org/cgit/ openstack/ gnocchi/ commit/ ?id=ee1740bf12c 13f546990f01ef7 6a4f0b29a78aeb
Committed: https:/
Submitter: Jenkins
Branch: stable/1.3
commit ee1740bf12c13f5 46990f01ef76a4f 0b29a78aeb
Author: Julien Danjou <email address hidden>
Date: Thu Jan 14 17:31:14 2016 +0100
ceph: fix the metric list to process with new measures
Currently, the list returned in the Ceph driver contains a lot of
doubloons because it returns a list and not a set. If 1 metric has N new
measures batch waiting to be processed, the list return will be size N
and not 1.
Using a set() avoids that issue and the memory draining implied.
Closes-Bug: #1533793 7a23a365babc1a2 537fb4d1052 a9d41bebc214014 e0c83a0e83)
Change-Id: I3a0b726aae14a1
(cherry picked from commit 614e13d47fdcaee