Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2019-01-08
16:50:01 sean-k-mooney stephenfin: oh you were refering to the cpu pinning bit
16:50:09 sean-k-mooney needing numa toplogy
16:50:10 jaypipes stephenfin: it's just the way you wrote that, made it seem that a request for dedicated CPU resources without any NUMA affinity/topology wasn't a valid request.
16:50:19 sean-k-mooney stephenfin: no its not
16:50:34 sean-k-mooney cpu pinning means i have a compute intensive workload
16:50:45 sean-k-mooney it dows nto mean that workload is memory sencitive
16:51:06 stephenfin jaypipes: Sorry, not what I was going for. If you are replying, to that email you might clarify
16:51:19 stephenfin sean-k-mooney: Think you're crossing wires here
16:51:36 stephenfin I'm saying oversubscription for pinned CPUs is daft, i.e. allowing two instances to share the same pCPU
16:51:36 jaypipes sean-k-mooney: pls see my comment on that topic on https://review.openstack.org/#/c/599957
16:52:03 jaypipes stephenfin: not a huge deal. if I reply, I'll mention it, but was more just making sure I was understanding you properly.
16:52:14 jaypipes stephenfin: see my comment on https://review.openstack.org/#/c/599957/ :)
16:52:15 stephenfin sean-k-mooney: I have no issues with unbinding CPU pinning from NUMA topologies
16:52:17 sean-k-mooney stephenfin: that i agree with but we were talking about memory oversubsription on the email tread
16:52:37 stephenfin jaypipes: Indeed :) I recall that discussion at the PTG
16:53:11 sean-k-mooney jaypipes: ya i stand by my comments on that spec i dont think we should support that
16:53:24 stephenfin sean-k-mooney: Right, you said "oversubscription is perfectly fine fore cpu pinning too" and I thought you mean CPU oversubscription
16:53:29 stephenfin *meant
16:53:33 sean-k-mooney jaypipes: that being https://review.openstack.org/#/c/599957/
16:53:57 stephenfin sean-k-mooney: Hence my "that is daft" comment :)
16:54:32 sean-k-mooney stephenfin: no dedicated means you get the whole cpus if we ever wanted to allow pinning and over subsrciption i woudl want to different cpu_policy to indicate that
16:55:09 openstackgerrit Merged openstack/nova stable/rocky: Make compute rpcapi version calculation check all cells https://review.openstack.org/624982
16:55:15 stephenfin sean-k-mooney: yup, I agree :) Just me misinterpreting what you said
16:55:21 sean-k-mooney we named it shared and dedicated to clearly callout that one allows over subscription or cpus and the other does not so https://review.openstack.org/#/c/599957/ is a definet no form me
16:56:20 sean-k-mooney stephenfin: i just resopned to that mail again pointing out the setting hw:numa_nodes=X does not prevent swaping the memory for 4k pages too incase you did not see
16:57:00 openstackgerrit Maciej Jozefczyk proposed openstack/nova master: Add fill_virtual_interface_list online_data_migration script https://review.openstack.org/614167
16:57:01 openstackgerrit Maciej Jozefczyk proposed openstack/nova master: Force refresh instance info_cache during heal https://review.openstack.org/591607
16:57:27 maciejjozefczyk jaypipes: I did a little update after your nits, will be silky smooth now
16:57:31 maciejjozefczyk needed to go, bb ;)
17:00:39 stephenfin sean-k-mooney, jaypipes, sahid: https://bugs.launchpad.net/nova/+bug/1810977
17:00:41 openstack Launchpad bug 1810977 in OpenStack Compute (nova) "Oversubscription broken for instances with NUMA topologies" [Undecided,New]
17:01:26 stephenfin I checked and reverting the patch gets us back to where we want to be. Have a better patch ready but currently trying to figure out why UTs won't run
17:01:45 jaypipes stephenfin: I would mark that bug as affecting me but... it doesn't. :)
17:02:37 sean-k-mooney im assuming oath does not over subscirbe ram or does not use the numa retlated features
17:03:28 jaypipes sean-k-mooney: that would be correct.
17:03:48 sean-k-mooney jaypipes: that certinly simplfies life.
17:04:09 jaypipes sean-k-mooney: indeed.
17:04:43 jaypipes sean-k-mooney: for performance-sensitive things, we use Ironic and the tenant does whatever the f**k they want to.
17:05:26 sean-k-mooney ya when thats a option baremetal will always beat even the most tunned vm
17:05:39 jaypipes (which is why 90% of our infrastructure is Ironic unfortunately, because tenants assume they must have baremetal to achieve the performance they need, even though 99% of our tenants haven't actually quantified what "performance" they "need".
17:05:52 openstackgerrit Merged openstack/nova-specs master: Add PENDING vm state https://review.openstack.org/554212
17:07:01 sean-k-mooney jaypipes: ya, im so familar with the tuneing options at this point that i pick an chooes the ones that make sense when i need them
17:07:34 sean-k-mooney i normally have at least one of the numa feature enabled even if it is just hw:numa_nodes=1
17:08:37 sean-k-mooney oh the pending state spec has merged cool :)
17:09:00 sean-k-mooney the rest of the premptable instance work is in the reaper project right?
17:15:16 jaypipes yonglihe: k, reviewed the NUMA spec.
17:51:01 openstackgerrit Merged openstack/nova-specs master: Flavor Extra Spec and Image Properties Validation https://review.openstack.org/618542
18:07:26 openstackgerrit Chris Dent proposed openstack/nova master: Remove link to placement configuration from nova config docs https://review.openstack.org/628185
18:07:26 openstackgerrit Chris Dent proposed openstack/nova master: Remove placement from nova install docs https://review.openstack.org/628184
18:07:27 openstackgerrit Chris Dent proposed openstack/nova master: Remove placement from contributor doc https://review.openstack.org/628186
18:07:45 openstackgerrit Chris Dent proposed openstack/nova master: Switch to using os-resource-classes https://review.openstack.org/628278
18:24:04 openstackgerrit Stephen Finucane proposed openstack/nova master: Fix overcommit for NUMA-based instances https://review.openstack.org/629281
18:24:13 stephenfin sean-k-mooney, gibi, jaypipes: Ta-da ^
18:24:22 stephenfin Makes sense to me, anyway. Lemme know
18:34:13 sean-k-mooney stephenfin: its almost right comment inline. you messed the case fo explcit small pages e.g. hw:mem_page_size=small|4|4k
18:40:31 stephenfin sean-k-mooney: replied
19:20:51 lbragstad johnthetubaguy jaypipes o/ curious if either of you have thoughts on this since the last time we discussed it - http://lists.openstack.org/pipermail/openstack-discuss/2019-January/001518.html
19:52:27 openstackgerrit Paul-Emile Element proposed openstack/os-traits master: Add new traits for TPM support https://review.openstack.org/629299
20:19:20 mnaser oh yikes
20:19:34 mnaser how safe is it to change the admin_or_owner rule
20:19:59 mnaser to say, something like: "is_admin:True or (role:compute and project_id:%(project_id)s)"
20:20:04 openstackgerrit Jack Ding proposed openstack/nova master: Improve libvirt image and snapshot handling https://review.openstack.org/616692
20:51:58 openstackgerrit Jack Ding proposed openstack/nova-specs master: Fix format flavor-extra-spec-image-property-validation https://review.openstack.org/629311
22:07:16 yonglihe jay, thanks.
22:27:17 yonglihe jaypipes, sean-k-mooney do we need aligned about sean-k-mooney's suggestion? move some stuff out of the 'nodes' and a more granular policy rule.
22:35:29 openstackgerrit Yongli He proposed openstack/nova-specs master: add spec "show-server-numa-topology" https://review.openstack.org/612256
22:37:38 yonglihe jaypipes, update per your comments. i working on a more specific policy rule, that's fine to me.
23:15:13 openstackgerrit Yongli He proposed openstack/nova-specs master: add spec "show-server-numa-topology" https://review.openstack.org/612256
23:15:19 jaypipes yonglihe: I actually think no, it's best just left defaulted to admin-only policy.
23:15:37 jaypipes yonglihe: but I won't hold up the spec for that reason or anything
23:16:05 yonglihe ok, i need revert it. already add a fine control
23:16:44 jaypipes yonglihe: don't revert :)
23:16:51 jaypipes yonglihe: it's not a big deal
23:16:54 yonglihe ok
23:18:04 yonglihe nice to know your idea about which item cause you add -1
23:18:32 yonglihe really nice hint for review.
23:40:33 openstackgerrit Merged openstack/python-novaclient stable/rocky: Fix up userdata argument to rebuild. https://review.openstack.org/613057
23:45:03 openstackgerrit Merged openstack/python-novaclient stable/rocky: Follow up "Fix up userdata argument to rebuild" https://review.openstack.org/613086
23:45:51 openstackgerrit Hongbin Lu proposed openstack/nova-specs master: Scheduling support for SR-IOV bond https://review.openstack.org/626055
#openstack-nova - 2019-01-09
00:03:54 openstackgerrit Hongbin Lu proposed openstack/nova-specs master: Scheduling support for SR-IOV bond https://review.openstack.org/626055
00:36:31 openstackgerrit Merged openstack/nova stable/rocky: Fail to live migration if instance has a NUMA topology https://review.openstack.org/625880
02:06:03 openstackgerrit Merged openstack/nova stable/rocky: Imported Translations from Zanata https://review.openstack.org/628109
02:06:09 openstackgerrit Merged openstack/nova stable/rocky: Fix a broken link https://review.openstack.org/626798
02:06:15 openstackgerrit Merged openstack/nova stable/rocky: Removing pip-missing-reqs from default tox jobs https://review.openstack.org/628952
02:06:24 openstackgerrit Merged openstack/nova stable/rocky: Fix destination_type attribute in the bdm_v2 documentation https://review.openstack.org/626874
02:25:59 takashin
04:07:05 openstackgerrit Takashi NATSUME proposed openstack/nova stable/queens: Add description of custom resource classes https://review.openstack.org/619125
04:07:10 openstackgerrit Takashi NATSUME proposed openstack/nova stable/queens: Add description of custom resource classes https://review.openstack.org/619125
07:38:32 openstackgerrit Zhenyu Zheng proposed openstack/nova-specs master: Amend the detach-boot-volume design https://review.openstack.org/619161
08:33:46 openstackgerrit Yikun Jiang proposed openstack/nova master: Per aggregate scheduling weight https://review.openstack.org/628163
08:58:43 gibi stephenfin: left a question in https://review.openstack.org/#/c/629281
09:23:51 sean-k-mooney gibi: i replied to your question. i think is a safe assumtion to make if they have enabled ram oversubcription that the smallest pagesize on the host is swappable and therefor can be over subsribed
09:24:34 sean-k-mooney the linux kernel does not allow you to reserve all memory on a node for hugepages so the smallest pagesize will always be the system default small page size
09:40:20 openstackgerrit Maciej Jozefczyk proposed openstack/nova master: Add fill_virtual_interface_list online_data_migration script https://review.openstack.org/614167
09:40:21 openstackgerrit Maciej Jozefczyk proposed openstack/nova master: Force refresh instance info_cache during heal https://review.openstack.org/591607
09:41:20 maciejjozefczyk sean-k-mooney: I found a bug in https://review.openstack.org/#/c/614167
09:42:09 gibi sean-k-mooney: thanks for the explanation
09:48:19 openstackgerrit Jan Gutter proposed openstack/nova master: Convert vrouter legacy plugging to os-vif https://review.openstack.org/571325
09:51:29 kashyap msg = _("Setting both 'live_migration_tunnelled' and"
09:51:29 kashyap Is this too verbose a message?

Earlier   Later