[Starlingx-discuss] memory reserved for vm not enough

Sun, Austin austin.sun at intel.com
Wed Jul 3 12:49:41 UTC 2019


Hi Yizhou:
You can use 'system host-memory-modify' command to adjust. memory config.
For example :
1) lock AIO controller-0
2) system host-memory-modify controller-0 <processor> -2M <the memory reserve for vswitch> -f vswitch.
3) unlock AIO controller-0.

When controller-0 is available again. the memory will be adjusted

Thanks.
BR
Austin Sun.

From: Xie, Cindy
Sent: Wednesday, July 3, 2019 8:37 PM
To: Xu, YizhouX <yizhoux.xu at intel.com>; 'starlingx-discuss at lists.starlingx.io' <starlingx-discuss at lists.starlingx.io>; Yang, Bin <bin.yang at intel.com>; Sun, Austin <austin.sun at intel.com>
Subject: RE: [Starlingx-discuss] memory reserved for vm not enough

Wondering if Austin or Bin can provide some insight here.

From: Xu, YizhouX [mailto:yizhoux.xu at intel.com]
Sent: Wednesday, July 3, 2019 7:40 PM
To: 'starlingx-discuss at lists.starlingx.io' <starlingx-discuss at lists.starlingx.io<mailto:starlingx-discuss at lists.starlingx.io>>
Subject: [Starlingx-discuss] memory reserved for vm not enough

Hi all:
               I'm testing pci-passthrough  with a  AIO stx node(iso version:20190607T142331Z), after install completed,
I  found that there's not enough memory reversed for my vms(server total memory is 64G,but only 7-8G can be allocated for vms),
Checked node with `system host-memory-list controller-0` and find that hugepages was configured and I did't deply dpdk-ovs.
Here are my questions:

1.      Did stx turn on hugepage as default? Can this feature be tured off to get all the reserved memory for my vms (I don't need dpdk-ovs)?

2.      If hugepage is necessary, refer to https://docs.openstack.org/nova/pike/admin/huge-pages.html ,I've Customized flavor for huge pages allocations with  `openstack flavor set m1.large --property hw:mem_page_size=large`
But it did't work (vm boot successfully , no memory allocated from reseved hugepages). Did I do it  in right way? Where can I trace the error?


Here 're the detail of my node:

[wrsroot at controller-0 ~(keystone_admin)]$ system host-memory-list controller-0
+-----------+---------+------------+---------+----------------+--------+--------+--------+-------+----------+--------+--------+----------+--------+--------+------------+---------------+
| processor | mem_tot | mem_platfo | mem_ava | hugepages(hp)_ | vs_hp_ | vs_hp_ | vs_hp_ | vs_hp | app_tota | app_hp | app_hp | app_hp_p | app_hp | app_hp | app_hp_pen | app_hp_use_1G |
|           | al(MiB) | rm(MiB)    | il(MiB) | configured     | size(M | total  | avail  | _reqd | l_4K     | _total | _avail | ending_2 | _total | _avail | ding_1G    |               |
|           |         |            |         |                | iB)    |        |        |       |          | _2M    | _2M    | M        | _1G    | _1G    |            |               |
+-----------+---------+------------+---------+----------------+--------+--------+--------+-------+----------+--------+--------+----------+--------+--------+------------+---------------+
| 0         | 51308   | 11000      | 51308   | True           | 1024   | 0      | 0      | None  | 1789952  | 22158  | 22158  | None     | 0      | 0      | None       | True          |
+-----------+---------+------------+---------+----------------+--------+--------+--------+-------+----------+--------+--------+----------+--------+--------+------------+---------------+

[wrsroot at controller-0 ~(keystone_admin)]$ cat /proc/meminfo | grep Huge
HugePages_Total:   22158
HugePages_Free:    22158
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB

Best Regards,
Xu, YiZhou









-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.starlingx.io/pipermail/starlingx-discuss/attachments/20190703/40c22b93/attachment.html>


More information about the Starlingx-discuss mailing list