Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-08-07
16:01:26 gibi dansmith: only one? ;)
16:01:52 dansmith gibi: many nodes, but only one in the openstack deployment :)
16:03:17 mnaser i'm going over the scheduler code and i've realized that we're pretty much sending instances to our biggest hypervisors and the smaller ones remain empty. the ram weighter by default weighs based on free memory, which means that a hypervisor with 128gb memory and all of it free will have less priority than one that has 384gb with 192gb free
16:03:39 mnaser arguably.. shouldnt it be based on a percentage?
16:04:03 mnaser https://github.com/openstack/nova/blob/master/nova/scheduler/weights/ram.py
16:04:19 jaypipes mnaser: perhaps use the num instances weigher instead?
16:04:43 dansmith or weigh the num_instances higher than the ram one
16:04:49 dansmith mnaser: some people want to pack first
16:04:50 jaypipes mnaser: nm, there is no such thing :(
16:04:52 dansmith because windows licenses
16:05:10 mnaser dansmith: packing would still work with -1.0
16:05:13 dansmith jaypipes: there's something like that pretty sure
16:05:29 mnaser https://github.com/openstack/nova/tree/master/nova/scheduler/weights
16:05:35 mnaser i dont see a num_instances on
16:05:38 mriedem stephenfin: where does this come from? https://review.openstack.org/#/c/490952/3/doc/source/admin/security-groups.rst
16:05:54 mnaser there's a fitler but not a weighter
16:05:58 stephenfin mriedem: cli-nova-manage-projects-security.rst
16:06:01 stephenfin (I think
16:06:03 mriedem oh nvm https://github.com/openstack/openstack-manuals/blob/stable/ocata/doc/admin-guide/source/cli-nova-manage-projects-security.rst
16:06:03 jaypipes mnaser: yeah :(
16:06:04 mriedem yeah
16:06:18 dansmith mnaser: yeah, that's what I'm thinking of, heh
16:06:47 mnaser cause if it was percentage based, the ability to pack vs distribute would still work, but would be based on usage percentage rather than just absolute free memory
16:07:24 mnaser i'm just slightly annoyed because i remember very clearly reading somewhere online (ML?) that someone had this exact issue and they dealt with it somehow but i figured "we won't deal with that" and here i am
16:08:17 dansmith man I really thought we had more weighers
16:09:40 mriedem stephenfin: so on https://review.openstack.org/#/c/490952/ I think we should have those various volume attach troubleshooting docs
16:10:18 jaypipes mnaser: can't believe I'm suggesting this, but you could adapt the existing RAM weigher and make it work on percentages instead of raw capacity numbers.
16:10:40 jaypipes mnaser: else submit a wishlist bug for one of us to get to.
16:10:57 jaypipes mnaser: like dansmith I'm kinda surprised such a thing doesn't already exist :(
16:11:57 mriedem stephenfin: and maybe leave todo comments for combining the quotas pages
16:13:15 mriedem stephenfin: if you want we can just deal with those comments in a follow up patch
16:14:02 stephenfin mriedem: Nah, I might as well address them there now
16:15:45 stephenfin mriedem: Also, what are your thoughts on https://review.openstack.org/#/c/453808 ? Can it still go in before RC1 if a FFE isn't required/suitable?
16:19:13 mriedem stephenfin: replied again
16:19:22 mriedem i don't see why --version is different from --local_cell
16:19:25 mriedem since they are both options
16:19:36 mriedem plus i feel like this is a distraction right now
16:21:56 mriedem sdague: how is https://review.openstack.org/#/c/490644/ different from https://review.openstack.org/#/c/478485/ ?
16:22:05 mriedem note i can't get to docs-draft.openstack.org/44/490644/3/check/gate-nova-docs-ubuntu-xenial/bd5e01d//doc/build/html/ right now
16:22:45 openstackgerrit Eric Fried proposed openstack/nova master: Get auth from context for glance endpoint https://review.openstack.org/490057
16:23:43 sdague mriedem: the TOC approach
16:24:10 sdague mriedem: see commit message
16:39:05 openstackgerrit Merged openstack/nova master: Add resource utilities to scheduler utils https://review.openstack.org/490514
16:44:01 mriedem sdague: do we need to worry about redirects for anything in https://review.openstack.org/#/c/490644/ ?
16:44:41 mriedem sdague: or is that handled by https://review.openstack.org/#/c/490644/3/doc/source/index.rst@186 ?
16:45:34 mnaser jaypipes do you think proposing a fix to replace the existing ram weighter by percentage would be an acceptable change
16:45:43 mnaser or introduce a new weighter instead?
16:45:59 mnaser because i genuinely think a ram weighter based on free memory only is not correct
16:47:08 mriedem sdague: nvm, looks like nothing actually moves
16:47:13 mriedem so no redirects needed
16:49:03 jaypipes mnaser: a new weigher would be better.
16:50:21 openstackgerrit Ed Leafe proposed openstack/nova master: Fix migrate single instance when it was created concurrently https://review.openstack.org/491439
16:51:26 edleafe bauzas: ^^ just fixed the pep8 issue
16:55:22 sdague mriedem: right, I did not move any source files
16:55:43 sdague I was just trying to build landing pages, and tame the TOC on the left sidebar which is a confusing mess right now
16:56:08 mnaser ehh
16:58:37 dansmith jaypipes: mnaser agreed
16:58:46 dansmith behavioral change may surprise people
16:59:10 mnaser yep, that's what i thought, i'm working on something now
16:59:15 jaypipes cool
16:59:48 stvnoyes mriedem mdbooth - I updated review https://review.openstack.org/#/c/389798/ - it's a small change. please take a look when you have a minute. Thanks.
17:01:16 jaypipes dansmith: correct.
17:16:46 stephenfin mriedem: Further replies left on https://review.openstack.org/#/c/453808/
17:17:04 stephenfin I know it's a distraction and would understand if it were ignored but it totally blocks the (undoubtedly fabulous) cliff integration for another cycle which would be rather depressing
17:18:26 openstackgerrit Merged openstack/nova master: Add Contributor Guide section page https://review.openstack.org/490644
17:18:30 mriedem stephenfin: why does --verbose block that but --local_cell does not?
17:18:47 mriedem why doesn't cliff allow us to have options, in other words
17:58:08 openstackgerrit Mohammed Naser proposed openstack/nova master: Introduce new percentage based RAM weigher https://review.openstack.org/491561
17:58:12 mnaser dansmith jaypipes i havent contributed to nova in a while so i welcome all critique whenever you have time ^
17:58:17 openstackgerrit Merged openstack/python-novaclient master: tools: Remove dead script https://review.openstack.org/480138
17:58:59 jaypipes mnaser: cheers
18:26:06 openstackgerrit Ed Leafe proposed openstack/nova master: Handle ironicclient failures in Ironic driver https://review.openstack.org/487925
18:37:40 jaypipes dansmith: what's the difference between rebuild and evacuate?
18:38:03 dansmith jaypipes: rebuild is in place on the same compute node
18:38:04 dansmith jaypipes: evacuate means "rebuild me somewhere else" and has to be run on a down compute node
18:38:12 dansmith jaypipes: http://www.danplanet.com/blog/2016/03/03/evacuate-in-nova-one-command-to-confuse-us-all/
18:38:33 jaypipes dansmith: thx. sorry, just needed a reminder after looking at the code... :(
18:38:45 dansmith np
18:44:45 mnaser dansmith jaypipes and evacuate when using ceph means start me somewhere else (not rebuilding)
18:45:07 jaypipes mnaser: when using ceph *for the root disk*. :)
18:45:08 dansmith mnaser: it's still a rebuild under the covers in nova,
18:45:14 dansmith it just doesn't erase the disk
18:45:17 dansmith which is covered in the above link
18:45:56 mnaser yes, that's correct. when reading that link, volume-backed made me think cinder but indeed
18:46:35 dansmith well, it says shared storage in one place too, but yeah
18:54:38 mriedem +2 on bauzas' IndexError fix https://review.openstack.org/#/c/491439/
18:58:18 sdague mriedem: also this doc page when you get a chance - https://review.openstack.org/#/c/490994/
18:59:03 sdague mriedem: +A
18:59:47 mriedem yup the tab is open
19:25:36 openstackgerrit Merged openstack/python-novaclient master: Updated from global requirements https://review.openstack.org/491296
19:26:22 openstackgerrit Matt Riedemann proposed openstack/nova master: Mark max microversion for Pike in history doc https://review.openstack.org/491581
19:26:22 openstackgerrit Matt Riedemann proposed openstack/nova master: Add a prelude section for Pike https://review.openstack.org/491424
19:26:23 openstackgerrit Matt Riedemann proposed openstack/nova master: Add release note for shared storage known issue https://review.openstack.org/491582
19:46:41 openstackgerrit Matt Riedemann proposed openstack/nova master: Add release note for shared storage known issue https://review.openstack.org/491582
19:49:09 openstackgerrit Merged openstack/python-novaclient master: Fix reservation_id not supported by Nova API https://review.openstack.org/468444
19:58:35 mriedem stephenfin: why didn't we migrate the hypervisor-specific config references? https://github.com/openstack/openstack-manuals/blob/stable/ocata/doc/config-reference/source/compute/hypervisor-hyper-v.rst
20:00:41 sdague mriedem: I think the answer to many things is going to be there are so many documents, and the code reviews are slow enough, that it's taking a while to get them in
20:01:20 mriedem i didn't know if it was intentional
20:01:47 sdague it's also about finding the right places for things to go
20:02:32 sdague yeh, I don't know, but I do know in building some of these sub pages that giving this thing any narative cohesion starting with 100 pages of mostly random notes people made at different depths, is challenging

Earlier   Later