Currently it's impossible to use 1Gb and 2Mb huge pages on one compute node
as a helper script from qemu package allows to use only one mount point
for huge pages, so 1 Gb huge pages support has to be managed explicitly.
This patch adds the following steps:
- create a mount point for 1 Gb huge pages and mount pages in runtime
- create a persistent record in /etc/fstab to mount huge pages on boot
- add related settings for libvirt
- update apparmor settings to allow an additional mount point
Due to LP #1560532 Nailgun doesn't provide information about allocated
huge pages when 1 Gb huge pages are used, so this information is computed
in a custom fact.
Closes-bug: #1566293
Change-Id: Iba421abdc354afa7d89f6f10a94c6ba3edb99148
(cherry picked from commit 10c3d88b55fbcd4427c4cb39690af9e0ae60175b)
Reviewed: https:/ /review. openstack. org/305741 /git.openstack. org/cgit/ openstack/ fuel-library/ commit/ ?id=c2b9eb32079 8357dc77557d638 f995254edea2a9
Committed: https:/
Submitter: Jenkins
Branch: stable/mitaka
commit c2b9eb320798357 dc77557d638f995 254edea2a9
Author: Sergey Kolekonov <email address hidden>
Date: Fri Apr 8 16:11:47 2016 +0300
Fix simultaneous usage of 1Gb and 2Mb huge pages
Currently it's impossible to use 1Gb and 2Mb huge pages on one compute node
as a helper script from qemu package allows to use only one mount point
for huge pages, so 1 Gb huge pages support has to be managed explicitly.
This patch adds the following steps:
- create a mount point for 1 Gb huge pages and mount pages in runtime
- create a persistent record in /etc/fstab to mount huge pages on boot
- add related settings for libvirt
- update apparmor settings to allow an additional mount point
Due to LP #1560532 Nailgun doesn't provide information about allocated
huge pages when 1 Gb huge pages are used, so this information is computed
in a custom fact.
Closes-bug: #1566293
Change-Id: Iba421abdc354af a7d89f6f10a94c6 ba3edb99148 427c4cb39690af9 e0ae60175b)
(cherry picked from commit 10c3d88b55fbcd4