Friday, 2020-10-09

« Thursday, 2020-10-08 Index Saturday, 2020-10-10 »
*** k_mouza has quit IRC00:04
*** freenzyfriday has joined #openstack-nova00:12
*** freenzyfriday has quit IRC00:16
*** k_mouza has joined #openstack-nova00:22
*** spatel has joined #openstack-nova00:23
*** k_mouza has quit IRC00:26
*** spatel has quit IRC00:28
*** k_mouza has joined #openstack-nova00:32
*** hamalq has quit IRC00:32
*** k_mouza has quit IRC00:36
*** LinPeiWen has joined #openstack-nova00:40
*** JamesBenson has quit IRC00:42
*** JamesBenson has joined #openstack-nova00:43
*** brinzhang has joined #openstack-nova00:44
*** JamesBenson has quit IRC00:47
*** yankcrime has quit IRC00:53
*** hemna has quit IRC00:56
*** hemna has joined #openstack-nova00:57
*** freenzyfriday has joined #openstack-nova01:10
*** zzzeek has quit IRC01:10
*** Liang__ has joined #openstack-nova01:10
*** zzzeek has joined #openstack-nova01:11
*** k_mouza has joined #openstack-nova01:13
*** freenzyfriday has quit IRC01:15
*** k_mouza has quit IRC01:17
*** JamesBenson has joined #openstack-nova01:20
*** JamesBenson has quit IRC01:24
*** k_mouza has joined #openstack-nova01:26
*** k_mouza has quit IRC01:30
*** LinPeiWen has quit IRC01:46
*** k_mouza has joined #openstack-nova01:57
*** k_mouza has quit IRC02:01
*** k_mouza has joined #openstack-nova02:23
*** k_mouza has quit IRC02:27
*** JamesBenson has joined #openstack-nova02:30
*** JamesBenson has quit IRC02:34
*** k_mouza has joined #openstack-nova02:51
*** freenzyfriday has joined #openstack-nova02:55
*** k_mouza has quit IRC02:55
*** LinPeiWen has joined #openstack-nova02:59
*** freenzyfriday has quit IRC03:04
*** JamesBenson has joined #openstack-nova03:05
*** psachin has joined #openstack-nova03:07
*** JamesBenson has quit IRC03:10
*** rcernin has quit IRC03:17
*** macz_ has joined #openstack-nova03:22
*** macz_ has quit IRC03:26
*** psachin has quit IRC03:40
*** psachin has joined #openstack-nova03:41
*** freenzyfriday has joined #openstack-nova03:49
*** freenzyfriday has quit IRC03:54
*** rcernin has joined #openstack-nova03:55
*** rcernin has quit IRC03:56
*** rcernin has joined #openstack-nova04:03
*** pmannidi has quit IRC04:16
*** mpsairam has joined #openstack-nova04:16
*** manoj_kumar_kata has joined #openstack-nova04:18
*** evrardjp has quit IRC04:33
*** evrardjp has joined #openstack-nova04:33
*** vishalmanchanda has joined #openstack-nova04:35
*** freenzyfriday has joined #openstack-nova04:43
*** freenzyfriday has quit IRC04:53
*** whoami-rajat__ has joined #openstack-nova05:02
*** zzzeek has quit IRC05:02
*** zzzeek has joined #openstack-nova05:05
*** ratailor has joined #openstack-nova05:07
*** manoj_kumar_kata has quit IRC05:25
*** freenzyfriday has joined #openstack-nova05:38
*** zzzeek has quit IRC05:42
*** zzzeek has joined #openstack-nova05:43
*** freenzyfriday has quit IRC05:43
*** rpittau|afk is now known as rpittau05:43
*** brinzhang_ has joined #openstack-nova05:52
*** brinzhang has quit IRC05:55
*** manoj_kumar_kata has joined #openstack-nova06:05
*** LinPeiWen has quit IRC06:11
*** zzzeek has quit IRC06:11
*** zzzeek has joined #openstack-nova06:12
*** LinPeiWen has joined #openstack-nova06:15
*** LinPeiWen has quit IRC06:23
*** freenzyfriday has joined #openstack-nova06:32
*** songwenping_ has joined #openstack-nova06:38
*** freenzyfriday has quit IRC06:41
*** songwenping__ has quit IRC06:42
*** tkajinam has quit IRC06:42
*** tkajinam has joined #openstack-nova06:42
*** freenzyfriday has joined #openstack-nova06:45
*** jobewan has quit IRC06:46
*** jobewan has joined #openstack-nova06:48
*** tesseract has joined #openstack-nova06:59
*** ralonsoh has joined #openstack-nova06:59
*** slaweq has joined #openstack-nova07:38
*** rcernin has quit IRC07:45
*** tosky has joined #openstack-nova07:46
*** sapd1 has joined #openstack-nova07:50
*** k_mouza has joined #openstack-nova08:20
*** whoami-rajat__ has quit IRC08:21
openstackgerritBalazs Gibizer proposed openstack/nova master: Ignore PCI devices with 32bit domain  https://review.opendev.org/75669708:26
*** derekh has joined #openstack-nova08:28
*** k_mouza_ has joined #openstack-nova08:28
*** k_mouza has quit IRC08:31
*** LinPeiWen has joined #openstack-nova08:36
*** dtantsur|afk is now known as dtantsur08:44
*** ociuhandu has joined #openstack-nova08:51
*** hemna_ has joined #openstack-nova09:00
*** hemna has quit IRC09:00
*** LinPeiWen has quit IRC09:05
lyarwoodstephenfin: image defined bdms09:06
lyarwoodstephenfin: block_device_mappings in the image metadata or something else?09:06
lyarwoodstephenfin: oh img_mappings09:06
lyarwoodstephenfin: I didn't think we used it anymore tbh09:09
lyarwoodstephenfin: I'll take a better look later today09:09
*** manoj_kumar_kata has quit IRC09:14
*** manoj_kumar_kata has joined #openstack-nova09:14
*** LinPeiWen has joined #openstack-nova09:28
*** yankcrime has joined #openstack-nova09:32
*** takamatsu has joined #openstack-nova09:37
*** manoj_kumar_kata has quit IRC09:40
*** manoj_kumar_kata has joined #openstack-nova09:40
openstackgerritwangzhh proposed openstack/nova master: Ignore InstanceNotFound exception when host rejoin cluster  https://review.opendev.org/75705309:41
*** Liang__ has quit IRC09:57
sean-k-mooneylyarwood: we use it here https://github.com/openstack/nova/blob/9fa563666e8500510def19269f740d99cab0e994/nova/compute/api.py#L833-L85610:21
sean-k-mooneywhich is used here https://github.com/openstack/nova/blob/9fa563666e8500510def19269f740d99cab0e994/nova/compute/api.py#L90710:23
lyarwoodsean-k-mooney: right we use block_device_mapping, I don't think we use img_mappings anywhere10:25
sean-k-mooneywe do10:25
sean-k-mooneyhttps://github.com/openstack/nova/blob/9fa563666e8500510def19269f740d99cab0e994/nova/compute/api.py#L933-L93410:25
sean-k-mooneywe merge the image mappings with the block device mappings10:26
sean-k-mooneybefore we append the flavor mappings10:26
sean-k-mooneyso right now we get the mappigns form the commandline the falvaor and iamge and merge them all together10:26
lyarwoodright but I still don't see any code looking up img_mappings10:27
sean-k-mooneythis does it https://github.com/openstack/nova/blob/9fa563666e8500510def19269f740d99cab0e994/nova/compute/api.py#L907-L90810:27
sean-k-mooneyhttps://github.com/openstack/nova/blob/9fa563666e8500510def19269f740d99cab0e994/nova/compute/api.py#L84110:27
lyarwoodsean-k-mooney: mappings != img_mappings10:28
sean-k-mooneyin the ovo we store "img_mappings" in the "mappings" field10:28
lyarwoodsean-k-mooney: that's my point10:28
sean-k-mooneylyarwood:yes it is10:28
sean-k-mooneyhttps://github.com/openstack/nova/blob/f24e054f5e41dcf45b7ecd20e6d41f4a293a91d2/nova/objects/image_meta.py#L55510:29
sean-k-mooneyimg_mappings is the legacy name for mappings10:29
sean-k-mooneywe coalese it and store it in mappings10:29
lyarwoodmappings is the legacy name10:30
sean-k-mooneyactully i guess its the other way around10:30
sean-k-mooneyok well the lookup should still work10:30
sean-k-mooneywe are using mappings which will retrun the same valuse as img_mappings10:30
sean-k-mooneybecause of https://github.com/openstack/nova/blob/f24e054f5e41dcf45b7ecd20e6d41f4a293a91d2/nova/objects/image_meta.py#L565-L57210:31
lyarwoodI guess the lookup is defined elsewhere?10:31
sean-k-mooneyactully that the set where is the get10:31
lyarwoodyeah10:32
sean-k-mooneyoh we just have it for set so when we build it we set it with the new name10:32
sean-k-mooneythe get is https://github.com/openstack/nova/blob/f24e054f5e41dcf45b7ecd20e6d41f4a293a91d2/nova/objects/image_meta.py#L657-L68510:33
lyarwoodbut https://github.com/openstack/nova/blob/9fa563666e8500510def19269f740d99cab0e994/nova/compute/api.py#L841 is specifically using the older name so that's always None10:33
sean-k-mooneyso is if from here https://github.com/openstack/nova/blob/9fa563666e8500510def19269f740d99cab0e994/nova/compute/api.py#L843-L84510:34
sean-k-mooneynot looking at the img_mappings would be a bug unless we had an explict deprecation of this fucntionaliy10:37
lyarwoodI'm still not entirely sure what it is the dict provides tbh, looking for an example10:38
lyarwoodvirtual image names to device names10:39
sean-k-mooneyi think you could pass anything you can pass in bdm on the commandline10:40
lyarwoodno that's stored in block_device_mapping or bdm_v210:41
lyarwoodthis is simple dict of things like {'root': '/dev/vda'} AFAICT10:41
lyarwoodhttps://github.com/openstack/nova/blob/4cf72ea6bfc58d33da894f248184c08c36055884/nova/block_device.py#L538-L58810:42
sean-k-mooneyi see in that case im less concerned10:42
sean-k-mooneywe dont provide stable/usersetable device names really anyway10:43
*** tbachman has quit IRC10:43
lyarwoodyup I wonder what this is actually used for tbh10:43
sean-k-mooneyhttps://github.com/openstack/nova/blob/f24e054f5e41dcf45b7ecd20e6d41f4a293a91d2/nova/objects/image_meta.py#L454-L46010:44
sean-k-mooneyit looks like it was never tested when we did the conversion10:46
sean-k-mooneyto ovos10:46
lyarwoodyeah we should get everything we need from img_block_device_mapping / block_device_mapping10:47
openstackgerritLee Yarwood proposed openstack/nova master: compute: Lock by instance.uuid lock during swap_volume  https://review.opendev.org/75469510:48
*** priteau has joined #openstack-nova10:50
*** rcernin has joined #openstack-nova10:50
*** zzzeek has quit IRC10:52
*** zzzeek has joined #openstack-nova10:54
*** zzzeek has quit IRC11:17
*** zzzeek has joined #openstack-nova11:18
openstackgerritLee Yarwood proposed openstack/nova master: Include removal of ephemeral backing files in the image cache manager  https://review.opendev.org/68942211:24
*** rcernin has quit IRC11:31
*** ratailor_ has joined #openstack-nova11:33
*** hemna_ has quit IRC11:33
*** hemna has joined #openstack-nova11:34
*** ratailor has quit IRC11:34
*** psachin has quit IRC11:34
*** ttx has quit IRC11:34
*** jamesdenton has quit IRC11:34
*** ttx has joined #openstack-nova11:36
*** scanepa has joined #openstack-nova11:38
*** psachin has joined #openstack-nova11:40
*** jamesdenton has joined #openstack-nova11:40
openstackgerritLee Yarwood proposed openstack/nova stable/ussuri: Use absolute path during qemu img rebase  https://review.opendev.org/75708311:58
openstackgerritLee Yarwood proposed openstack/nova stable/train: Use absolute path during qemu img rebase  https://review.opendev.org/75708411:59
openstackgerritLee Yarwood proposed openstack/nova stable/stein: Use absolute path during qemu img rebase  https://review.opendev.org/75708511:59
*** sapd1 has quit IRC12:01
*** slaweq has quit IRC12:03
*** slaweq has joined #openstack-nova12:17
*** freenzyfriday has quit IRC12:26
*** psachin has quit IRC12:31
openstackgerritStephen Finucane proposed openstack/nova master: docs: Add image metadata property reference guide  https://review.opendev.org/75686712:48
stephenfinlyarwood: Context on that img_mappings question was ^ I was trying to document what it was for and remove it if it wasn't necessary. I resorted to marking it as internal only and kicking the can down the road :-)12:49
stephenfinthough it sounds like you and sean-k-mooney might have teased it out, /me reads back up12:49
lyarwoodstephenfin: yeah tbh I don't think it's being used given what we were talking about above12:50
lyarwoodbrb12:50
stephenfinsean-k-mooney: I replied to your questions on that doc review. tl;dr: auto generating this stuff is too hard, and I'm not looking to replace Glance's XML schemas for code-based verification purposes12:51
sean-k-mooneystephenfin:ya looking at the code we still have code to use the legacy name12:53
sean-k-mooneyhowever the data is nologer stored usign the legacy name so that code is dead12:53
sean-k-mooneyso i dont think img_mappins is used and we just use the image bdms12:54
sean-k-mooneystephenfin: and ya i think if doc strings are not easy to use i guess doing it manually is ok12:54
sean-k-mooneyit would be nice if they could not get out of sync12:55
stephenfinAgreed, but I just don't think it's practical in this case unfortunately12:55
sean-k-mooneyi would still like to encurage peopel to update glance whenever tehy add new image metadata12:55
sean-k-mooneyoh by the ay i also didnt repond to the pinning patch12:56
sean-k-mooneyyes i forgot we reused cpuset to avoid modifing the object12:56
sean-k-mooneyill proably remove my -1 later just finishing up a bz comment beofre the bug call so havent got to reviews yet today12:57
stephenfincool12:57
sean-k-mooneyi saw your latest commnt in email however12:57
* sean-k-mooney actully read my gerrit emails regurally12:58
*** rpittau is now known as rpittau|afk13:03
*** ratailor_ has quit IRC13:03
*** ratailor has joined #openstack-nova13:03
lyarwoodelod: huh weird, I was sure this landed in stable/victoria13:09
lyarwoodelod: thanks for catching that13:09
openstackgerritLee Yarwood proposed openstack/nova stable/victoria: Use absolute path during qemu img rebase  https://review.opendev.org/75709813:10
openstackgerritLee Yarwood proposed openstack/nova stable/ussuri: Use absolute path during qemu img rebase  https://review.opendev.org/75708313:10
openstackgerritLee Yarwood proposed openstack/nova stable/train: Use absolute path during qemu img rebase  https://review.opendev.org/75708413:10
openstackgerritLee Yarwood proposed openstack/nova stable/stein: Use absolute path during qemu img rebase  https://review.opendev.org/75708513:10
*** nweinber has joined #openstack-nova13:11
*** ratailor has quit IRC13:15
elodlyarwood: np :)13:16
CeeMacafternoon13:28
CeeMacquick question, is there a way to set the libvirt cpu_mode using metadata or some such on a flavour/image/instance at all?13:28
sean-k-mooneystephenfin: how would you feel about default all instance with a numa toplogy to hw:mem_page_size=small unless otherwise set. this is needed for numa in placment anyway but would you be open to doing this anyway even without numa in placment13:28
sean-k-mooneywe might want to restrict it to new instance only but we do have customer hit this edgecase a lot13:29
sean-k-mooneythe set hw:cpu_policy=dediecated and dont realise that unless they also set hw:mem_page_size=small then there vm will eventually get killed by the OOM reaper13:30
sean-k-mooneysince the vms memory is pined to a numa node but the did nto enable per numa memory tracking13:31
sean-k-mooneystephenfin:by the way for your image properties patch should we send a mail to the list to let the glance folks know we are going to move documenation of image properties back into nova13:33
stephenfinsean-k-mooney: I don't think an email is necessary. I'll just propose a patch to glance to remove the existing doc and point back to nova. If they want to keep their copy, that's fine13:43
stephenfinsean-k-mooney: As for mem pages, I'm not sure. You can't overcommit with 'hw:mem_page_size' set, right? Even if you could, your overcommit would be against memory from that cell, rather than the entire host. Not necessarily a bad thing but both things are changes in behavior13:44
sean-k-mooneyright now it broken to not set it13:47
sean-k-mooneywe will use the global memory when determining if the vm wil fit then pin it to one numa node13:47
sean-k-mooneyso over commit is broken in any case13:47
sean-k-mooneyonce we track numa in placment its not negoicable it has to do numa aware memroy tracking which is what setting that would do13:48
sean-k-mooneywe have had 2-3 customer hit this in the last year or so13:48
sean-k-mooneyfor numa in placment this is adress by partioning the cloud into numa nost and non numa hosts.13:51
sean-k-mooneywe currently dont enforce that you se hw:mem_page_size if you have a numa toplogy but its always a bug/user errror to not set it13:52
stephenfinif we can figure out a way to do overcommit on a per-node basis, then I'd be okay with that13:55
stephenfinotherwise, it's a big change13:55
sean-k-mooneywell as is said overcomit does not work properly today13:56
sean-k-mooneythe OOM reaper will kill the vm if the numa node runs out of memory13:56
sean-k-mooneyeven if the othe rnuma node is fully empty13:56
stephenfinyou could say the same thing about overcommit without NUMA13:57
sean-k-mooneyits not the same13:57
stephenfinno, but it's similar13:57
sean-k-mooneyif you set over commit to 1.0 e.g. no over commit13:57
sean-k-mooneywith the current state the vms can be killed becuase the numa node is exausted13:58
sean-k-mooneyunless you set hw:mem_page_size=small13:58
stephenfinyes, I understand that13:59
sean-k-mooneyits requried for numa in plamcent for numa hosts13:59
sean-k-mooneyfor non numa hosts i was not going to be an issue since numa instance could not land there13:59
stephenfinbut if you had work and instances evenly distributed between NUMA nodes, then it would behave the same as the non-NUMA case13:59
sean-k-mooneyso we defered this conversation the last 2 cycles because numa in palcment will prevent it form happening13:59
sean-k-mooneystephenfin: without my numa loadbancing feature it wont evenly disturbute14:00
sean-k-mooneyunless you consume all the cpus on the first node14:00
sean-k-mooneyeven then it won evenly disturbute14:00
openstackgerritDan Smith proposed openstack/nova master: Allow excluding image type capabilities  https://review.opendev.org/75653414:02
openstackgerritBalazs Gibizer proposed openstack/nova master: Refactor _claim_pci_device_for_interface_attach to prepare for qos  https://review.opendev.org/75689514:03
openstackgerritBalazs Gibizer proposed openstack/nova master: Allow extending an existing instance allocation  https://review.opendev.org/75710914:03
openstackgerritBalazs Gibizer proposed openstack/nova master: Make remove allocation symmetric with add allocation  https://review.opendev.org/75711014:03
openstackgerritBalazs Gibizer proposed openstack/nova master: Factor out port resource deallocation  https://review.opendev.org/75711114:03
openstackgerritBalazs Gibizer proposed openstack/nova master: Support interface attach with qos ports  https://review.opendev.org/75653014:07
*** ociuhandu has quit IRC14:11
*** ociuhandu has joined #openstack-nova14:12
*** k_mouza_ has quit IRC14:25
*** k_mouza has joined #openstack-nova14:26
CeeMacany thoughts on if its possible to set the libvirt cpu_mode value against an image, flavour or instance?14:40
*** slaweq has quit IRC14:42
*** lbragstad_ is now known as lbragstad14:49
*** freenzyfriday has joined #openstack-nova14:51
*** k_mouza has quit IRC14:52
*** freenzyfriday has quit IRC14:55
*** mlavalle has joined #openstack-nova14:57
sean-k-mooneyCeeMac: its not possible and it was reject in both the image and flaovr in the past15:03
*** k_mouza has joined #openstack-nova15:04
sean-k-mooneyCeeMac: we would prefer you to use traits to sate what instuctions you need instead15:04
sean-k-mooneywe will then schdule you to a host that has a cpu model that you can use15:04
sean-k-mooneycpu_mode specificly is defined by the admin in the config and not overrideable because it has live migration implciations15:05
CeeMacsean-k-mooney: thanks, I'm trying to expose hardware virtualization to a specific instance and didn't want to go down the line of making a global cpu_mode change15:06
CeeMacsean-k-mooney: do you know which / if any specific traits would be avaialable to do that?15:06
sean-k-mooneyCeeMac: we intentionally dont expose this via a user or public api as its a interoperatblity issue and only applys to one virt dirver15:06
sean-k-mooneyCeeMac: do you mean nested virt e.g. expose vmx15:07
CeeMacyes15:07
CeeMackvm is enabled for nested15:07
sean-k-mooneyor are you asking to chdule based on qemu vs kvm15:07
sean-k-mooneyah15:07
CeeMacbut the guest isn't seeing the flag15:07
sean-k-mooneythen you can use vmx i bleive15:07
sean-k-mooneyhttps://github.com/openstack/os-traits/blob/master/os_traits/hw/cpu/x86/intel.py#L2615:08
sean-k-mooneyso HW_CPU_X86_VMX15:08
sean-k-mooneysorry15:08
sean-k-mooneyHW_CPU_X86_INTEL_VMX15:09
sean-k-mooneyfor amd systems you would use HW_CPU_X86_SVM15:09
sean-k-mooneywe dont currently have a way to say X or Y unfutetly15:09
CeeMacthanks, is that something I can specify directly on an instance? or is it added at the image or flavour?15:10
sean-k-mooneyno you can only set the required trait in the flaovr/image15:10
sean-k-mooneynot per instnace15:10
CeeMacright15:10
sean-k-mooneydoign things per instance does not really work with nova15:10
CeeMacso its flavour or image?15:10
sean-k-mooneyyes both work15:10
CeeMacyeah i noticed that :)15:10
CeeMaccool15:10
CeeMacgreat thanks a lot15:10
CeeMaci'll try creatign a custom flavour15:11
sean-k-mooneyCeeMac: that will just ensure you land on a host that has nested vert enabled15:12
sean-k-mooneybut all other vms on that host will also get it15:12
sean-k-mooneyconfigureing nested virt is a system wide thing15:13
CeeMacright15:13
sean-k-mooneythe other way that people do this without a trait15:13
CeeMacso the vmx flag should automatically be exposed to the guest if KVM is configured for nested?15:13
sean-k-mooneyis to use the instance extra and host aggrates15:14
sean-k-mooneyam yes if you use host model or host-passthough as the mode15:14
*** songwenping__ has joined #openstack-nova15:14
melwittdansmith: I found a small bug in the cherry-pick hash check script, if you wouldn't mind reviewing https://review.opendev.org/75663915:15
dansmithmelwitt: cool15:16
melwittthanks!15:16
CeeMaci don't have cpu_mode set in nova.conf by the look of it15:16
*** songwenping_ has quit IRC15:17
CeeMacand looking at the virsh dumpxml for an instance the cpu mode is 'custom' with a fallback model Skylake-Client-IBRS15:17
sean-k-mooneyCeeMac if you dont set it it default to None which in the libvirt driver is converted to host-model15:18
melwittlyarwood: easy fix for the cherry-pick hash check script if you get a moment https://review.opendev.org/75663915:18
*** lpetrut has joined #openstack-nova15:19
CeeMacsean-k-mooney: thanks, i think I recall reading something like that15:19
CeeMaci can't seem to find where to set the trait in metadata15:19
CeeMaccould be veing blind though15:19
sean-k-mooneyopenstack flavor set --property trait:HW_CPU_X86_VMX=retuired <flavor>15:20
sean-k-mooneyi think that is the correct command15:21
sean-k-mooneystephenfin: ^15:21
sean-k-mooneystephenfin: do you know if you have to quote the trait?15:21
CeeMacah yes, the old cli :)15:21
sean-k-mooney... retuired should be required15:21
CeeMacgot it15:21
sean-k-mooneyCeeMac: never trust my spelling15:22
stephenfinsean-k-mooney: only if there are spaces or things the shell could confuse15:22
stephenfinI usually quote to be safe15:22
sean-k-mooneyya i often do but i was not sure if it would mess up the parsing of the =15:22
CeeMacstephenfin: so "trait:.....required" <flavour>15:24
stephenfinCeeMac: Yup. We list the supported versions here https://docs.openstack.org/nova/latest/configuration/extra-specs.html#trait15:25
stephenfin*supported traits15:25
CeeMacperfect, thanks15:25
sean-k-mooneyCeeMac: just be aware that we and them together15:26
sean-k-mooneyso if you use the amd trait and the intel trait then it wont find a host15:26
CeeMacright, good to know :)15:26
sean-k-mooneyif you have a mix you should use a custom extra spec instead of a trait and use host aggreates15:26
CeeMaci'm pretty sure at this point I don't want to do it at all, but i'll try it in my test env to be sure15:27
CeeMacstephenfin: from that link it looks like I need trait{group}:HW_CPU_X86_VMX15:27
CeeMacis the {group} specific to something?15:27
sean-k-mooneythere was a proposal to support "in" so you coudl do somehtin like trait:in(intel...,amd...)=required15:28
sean-k-mooneyCeeMac: you shoudl avoid groups15:28
CeeMaci can do that15:28
sean-k-mooneyjust leave out the group15:28
CeeMacgot it15:28
sean-k-mooneywe allow named or number grousp of triats15:28
sean-k-mooneybut they are hard to use and we may eventually remove it form the flaovr15:28
*** k_mouza has quit IRC15:28
CeeMacyeah, i'm not a  big fan of "hard to use" so I'm ok with that :D15:29
sean-k-mooneygroups dont interact well when you have request form multipel sources15:29
sean-k-mooneye.g. flavor, neutron and cyborg15:29
sean-k-mooneythere are some other gotchas so  unless you know how its implmented you should avoid them15:30
CeeMaci'll avoid them15:33
CeeMacto be honest, i'd rather not  allow nested virtualisation, but I'm being instructed to see if it is an option, so yeah15:33
*** k_mouza has joined #openstack-nova15:35
*** jangutter_ has joined #openstack-nova15:35
*** lpetrut has quit IRC15:35
*** freenzyfriday has joined #openstack-nova15:37
*** jangutter has quit IRC15:39
sean-k-mooneyCeeMac: i always use nested vert and have done for years but its not for everyone15:39
sean-k-mooneyit has less sharp edges then it used too15:39
*** k_mouza has quit IRC15:39
CeeMaclooks like OSA is set up to support it out of the box as well from a kernel module perspective15:40
sean-k-mooneyCeeMac:it was enabled by default in the upstream 4.19 kernel15:40
CeeMacif it works without me having to make any major changes and reboot everything then I'm ok with that15:41
CeeMacoh15:41
*** macz_ has joined #openstack-nova15:41
CeeMacfair enough :)15:41
*** k_mouza has joined #openstack-nova15:41
CeeMacinstance has deployed using  a flavor with that trait15:41
CeeMacso looking fine so far15:41
CeeMacits more of a legacy issue with my user base than a technical objection15:42
*** sapd1 has joined #openstack-nova15:42
CeeMaca lot of them still don't understand that IaaS means they're on their own unless they want to pay for support services15:43
sean-k-mooneyya15:45
sean-k-mooneythe old issue wit nested virt was you could not migrate the l1 vm if the l1 guest was running level 2 guests15:46
CeeMacah, yeah i could see that being a problem15:46
sean-k-mooneyits been fix in 4.1915:47
*** freenzyfriday has quit IRC15:47
sean-k-mooneythat was on of the technical reason operator did not often use it in the past15:47
sean-k-mooneysince it limited how they could use live migration15:48
CeeMacmakes sense15:50
*** k_mouza has quit IRC15:55
*** k_mouza has joined #openstack-nova15:57
openstackgerritStephen Finucane proposed openstack/nova master: libvirt: Only add a USB controller if it's necessary  https://review.opendev.org/75654915:58
openstackgerritStephen Finucane proposed openstack/nova master: libvirt: Remove support for '[libvirt] use_usb_tablet'  https://review.opendev.org/75655015:58
openstackgerritStephen Finucane proposed openstack/nova master: libvirt: Rationalize attachment of USB tablet  https://review.opendev.org/75655115:58
openstackgerritStephen Finucane proposed openstack/nova master: libvirt: Add support for virtio-based input devices  https://review.opendev.org/75655215:58
sean-k-mooneystephenfin: did you add None to the pointer device. the original requirement diring this is to have no input device attached to the guest15:59
sean-k-mooneyso using virtio wont actully help in the realtime case16:00
sean-k-mooneyi assume that is just a nice to have while your here16:00
*** tesseract has quit IRC16:01
admin0hi guys .. can the default domain be null or "" blank ?16:03
admin0insetad of .novalocal ?16:03
sean-k-mooneydns domain16:03
admin0can it be null ?16:03
admin0or some value has to be there16:03
sean-k-mooneyits set by https://docs.openstack.org/nova/latest/configuration/config.html#api.dhcp_domain16:04
admin0it does not say if it can be left null16:04
sean-k-mooneyi dont think it can be16:05
sean-k-mooneyalthough we have exampel of it being empty16:05
*** priteau has quit IRC16:06
sean-k-mooneywe use it like this16:06
sean-k-mooneyhttps://github.com/openstack/nova/blob/58636d34ec68d44952b22947bac01d4b88ae103a/nova/api/metadata/base.py#L533-L53616:06
*** k_mouza has quit IRC16:06
sean-k-mooneyyou would have to set it to ""16:06
sean-k-mooneyhttps://github.com/openstack/nova/blob/85c35b4d9d6386866ddf197e4a751f729b726188/nova/conf/api.py#L218-L22216:06
sean-k-mooneybut i dont know if that wil work or not16:07
sean-k-mooneydefault might still replace ""16:07
sean-k-mooneybut it would be easy to test16:07
sean-k-mooneyjust set [api]/dhcp_domain=""16:07
*** manoj_kumar_kata has quit IRC16:10
melwittit can be blank. we had it set that way at yahoo16:12
*** links has joined #openstack-nova16:16
*** ociuhandu has quit IRC16:21
*** links has quit IRC16:25
*** dtantsur is now known as dtantsur|afk16:26
*** raildo has quit IRC16:30
*** raildo has joined #openstack-nova16:30
*** gyee has joined #openstack-nova16:32
stephenfinsean-k-mooney: No, it wasn't helpful. You can't disable the built-in PS2 devices16:35
*** tosky has quit IRC16:35
stephenfinHowever16:35
sean-k-mooneyyou cant by generating model=none in the xml16:35
stephenfinI don't think that's supported. The only RFE I found for it was DEFERRED16:36
sean-k-mooneywell the perfromacne team teh filed teh orignal downstream issue had said they disabeld it manually by editing the xml16:36
stephenfinthey were talking about USB devices16:37
sean-k-mooneyyes16:37
stephenfinyou won't get a USB keyboard or pointer device on x86 unless you request them via image metadata properties or nova.conf16:37
sean-k-mooneyno libvirt will add it16:37
stephenfinso all that's necessary is to remove the superfluous USB bus16:37
stephenfinnot a USB one, it won't16:37
stephenfinit'll add a PS2 mouse and keyboard16:38
stephenfinnot USB16:38
sean-k-mooneyok but i think they wanted to remvoe that too16:38
stephenfinSee above. I don't think that's supported16:38
sean-k-mooneyya so your proposing what use virtio?16:39
sean-k-mooneyor ps216:39
stephenfinnope16:39
stephenfinI've four patches16:39
sean-k-mooneythey are hoping to remove all interupts related to the input device16:39
stephenfinfirst one is to remove the unnecessary USB bus16:39
stephenfinWith that, you can set '[compute] pointer_model' to 'none' or 'ps2mouse' and, unset the 'hw_pointer_model' image metadata property prop, and ensure you haven't set 'hw_*_bus' to 'usb'16:40
stephenfinand you won't have any USB devices or a USB bus16:41
stephenfinthe second is a cleanup patch16:41
sean-k-mooneywell we dont want this to be in the nova.conf really16:41
sean-k-mooneythat will cause issue for live migation16:41
stephenfinthe third makes it so that we only attach a USB tablet if graphics (SPICE or VNC) are enabled16:41
stephenfinbecause it's useless otherwise16:41
*** priteau has joined #openstack-nova16:42
stephenfinand the fourth adds support for virtio devices because they're supposed to be far more performant and it was an easy win16:42
stephenfinAgreed. However, I can't remove it until you add that system metadata thing for Q3516:43
sean-k-mooney:( ya16:43
sean-k-mooneyok we might want to update the downstream bz to say the input device cant be removed but we are goign to do this instead16:44
stephenfinI don't think the PS2 device is a massive issue. It's the USB bus that was doing most of the damage16:44
stephenfinBut I should check again to be sure16:45
stephenfinMonday! For now, it's sunny outside so o/ :)16:45
sean-k-mooneyi guess thre expect result was actully "no USB and no graphics"16:45
*** hamalq has joined #openstack-nova16:45
sean-k-mooneyso ps2 device or virtio might be fine actully16:46
sean-k-mooneybased on https://bugzilla.redhat.com/show_bug.cgi?id=178744216:46
openstackbugzilla.redhat.com bug 1787442 in documentation ""Configuring Real-Time Compute" lacks guest device configuration tuning" [Unspecified,Closed: wontfix] - Assigned to stephenfin16:46
*** ViniciusCoelho has joined #openstack-nova16:46
*** hemna has quit IRC16:50
admin0sean-k-mooney, i made it blank and it still gives me a .novalocal16:50
admin0is that something from neutron in dhcp ?16:51
*** hemna has joined #openstack-nova16:51
sean-k-mooneyno i think since novalocal is the default of that coning option you have to do =""16:51
sean-k-mooneyi dont think blank will actully work16:51
sean-k-mooneymelwitt: do you recal if ye set dhcp_domain= or dhcp_domain=""16:52
melwittdhcp_domain=""16:52
melwittset to empty string16:53
melwittsorry, that's what I meant when I said "blank" earlier16:53
sean-k-mooneyya ok that makes sense16:55
sean-k-mooneydhcp_domain= i think will be treated like its unset and the default will be used16:55
*** derekh has quit IRC16:55
sean-k-mooneyadmin0: this might only take effect for new vms too16:56
*** links has joined #openstack-nova16:56
admin0unless i can modify the dhcp config manually16:56
sean-k-mooneyfor existing vm i think we store there hostname in the db16:56
sean-k-mooneyyou would need to update the instnace.hostname filed in the db16:57
sean-k-mooneyhttps://github.com/openstack/nova/blob/master/nova/objects/instance.py#L13016:57
*** hemna has quit IRC17:10
*** hemna has joined #openstack-nova17:12
*** sapd1 has quit IRC17:13
*** hemna has quit IRC17:44
*** hemna has joined #openstack-nova17:49
*** hack-char has quit IRC17:50
*** hack-char has joined #openstack-nova17:50
*** priteau has quit IRC17:58
openstackgerritmelanie witt proposed openstack/nova stable/queens: [stable-only] Add functional test for bug 1731668  https://review.opendev.org/75663618:00
openstackbug 1731668 in OpenStack Compute (nova) queens "placement: claim allocations fails with IndexError in _ensure_lookup_table_entry" [Low,In progress] https://launchpad.net/bugs/1731668 - Assigned to melanie witt (melwitt)18:00
openstackgerritmelanie witt proposed openstack/nova stable/queens: [stable-only] Use a separate transaction for reading after race  https://review.opendev.org/75663718:00
openstackgerritmelanie witt proposed openstack/nova stable/queens: [stable-only] Add functional test for bug 1731668  https://review.opendev.org/75663618:03
openstackbug 1731668 in OpenStack Compute (nova) queens "placement: claim allocations fails with IndexError in _ensure_lookup_table_entry" [Low,In progress] https://launchpad.net/bugs/1731668 - Assigned to melanie witt (melwitt)18:03
openstackgerritmelanie witt proposed openstack/nova stable/queens: [stable-only] Use a separate transaction for reading after race  https://review.opendev.org/75663718:03
melwittomg uuughh18:03
openstackgerritmelanie witt proposed openstack/nova stable/queens: [stable-only] Add functional test for bug 1731668  https://review.opendev.org/75663618:05
openstackbug 1731668 in OpenStack Compute (nova) queens "placement: claim allocations fails with IndexError in _ensure_lookup_table_entry" [Low,In progress] https://launchpad.net/bugs/1731668 - Assigned to melanie witt (melwitt)18:05
openstackgerritmelanie witt proposed openstack/nova stable/queens: [stable-only] Use a separate transaction for reading after race  https://review.opendev.org/75663718:05
melwittsorry for the spam18:06
*** ralonsoh has quit IRC18:10
*** vishalmanchanda has quit IRC18:15
*** tkajinam has quit IRC18:25
*** hemna has quit IRC18:26
*** tosky has joined #openstack-nova18:32
*** k_mouza has joined #openstack-nova18:32
*** links has quit IRC18:41
*** priteau has joined #openstack-nova18:42
*** priteau has quit IRC18:48
openstackgerritmelanie witt proposed openstack/nova stable/train: Add a workaround config toggle to refuse ceph image upload  https://review.opendev.org/75717719:05
*** priteau has joined #openstack-nova19:14
*** manoj_kumar_kata has joined #openstack-nova19:20
*** k_mouza has quit IRC19:20
*** ociuhandu has joined #openstack-nova19:24
*** priteau has quit IRC19:27
melwittmnaser: is ^ something you would find useful/appropriate for sta19:47
melwitt*stable/train or nah?19:47
*** zzzeek has quit IRC19:49
*** ociuhandu has quit IRC19:50
*** manoj_kumar_kata has quit IRC19:50
*** zzzeek has joined #openstack-nova19:51
*** hemna has joined #openstack-nova20:20
*** hemna has quit IRC20:32
*** hemna has joined #openstack-nova20:32
*** nweinber has quit IRC20:37
*** ociuhandu has joined #openstack-nova20:51
*** ociuhandu has quit IRC21:13
*** ViniciusCoelho has quit IRC21:18
*** ociuhandu has joined #openstack-nova21:18
*** ociuhandu has quit IRC21:25
*** raildo has quit IRC21:30
*** freenzyfriday has joined #openstack-nova21:32
*** freenzyfriday has quit IRC21:37
*** hamalq has quit IRC21:37
*** artom has quit IRC21:51
*** hamalq has joined #openstack-nova22:00
*** freenzyfriday has joined #openstack-nova22:26
*** takamatsu has quit IRC22:26
*** freenzyfriday has quit IRC22:35
*** macz_ has quit IRC23:06
*** freenzyfriday has joined #openstack-nova23:07
*** freenzyfriday has quit IRC23:12
*** mlavalle has quit IRC23:19
*** hamalq has quit IRC23:26
*** tosky has quit IRC23:26
*** songwenping__ has quit IRC23:30
*** songwenping__ has joined #openstack-nova23:30
« Thursday, 2020-10-08 Index Saturday, 2020-10-10 »

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!