diff options
| author | Wangpan <hzwangpan@corp.netease.com> | 2013-01-16 10:00:47 +0800 |
|---|---|---|
| committer | Wangpan <hzwangpan@corp.netease.com> | 2013-01-18 13:05:44 +0800 |
| commit | 796716d4cc51f35a734aa1f9d70ea567c37b5a79 (patch) | |
| tree | b0d7c675d9fe258d8a9809992289d84dec1607f9 /nova/exception.py | |
| parent | e3a729b7c8873146d00d915a07094d327f97d184 (diff) | |
| download | nova-796716d4cc51f35a734aa1f9d70ea567c37b5a79.tar.gz nova-796716d4cc51f35a734aa1f9d70ea567c37b5a79.tar.xz nova-796716d4cc51f35a734aa1f9d70ea567c37b5a79.zip | |
Fix wrong quota reservation when deleting resizing instance
When we delete an instance which is resizing from m1.small to m1.tiny, we use
instance['memory_mb'] to create quota's ram reservation, if the resizing is not
'finished', the in_use value in quota_usages is NOT commited(2048M),
but the instance['memory_mb'] is updated(512M) in _finish_resize(), so the
deleting operation after updating will commit a wrong ram reservation(-512M),
and the in_use value changes to 2048-512=1536.
The cores reservation encounters the same problem.
Fixes bug #1099729
Change-Id: Ide54fc9d822041f535819c496822aeac5a8dc68d
Diffstat (limited to 'nova/exception.py')
0 files changed, 0 insertions, 0 deletions