Friday, 2017-09-29

*** baoli has quit IRC00:00
*** Swami has quit IRC00:01
*** smatzek has quit IRC00:01
*** baoli has joined #openstack-nova00:01
*** penick has quit IRC00:04
*** catintheroof has quit IRC00:06
*** mingyu has quit IRC00:11
*** mingyu has joined #openstack-nova00:14
*** gbarros has joined #openstack-nova00:15
*** thorst has quit IRC00:15
*** gbarros has quit IRC00:17
*** yangyapeng has joined #openstack-nova00:18
*** yangyapeng has quit IRC00:23
*** alex_xu has joined #openstack-nova00:25
*** alex_xu has quit IRC00:25
*** alex_xu has joined #openstack-nova00:25
*** yingjun has joined #openstack-nova00:26
*** thorst has joined #openstack-nova00:26
*** gbarros has joined #openstack-nova00:28
*** rcernin has quit IRC00:29
sean-k-mooneydansmith: qq im setting up the live migration environmnet for the multi port binding work00:30
sean-k-mooneydevstack complete sucessfully but i dont think i registerd the hyperviors correctly in the toplevel api db http://paste.openstack.org/show/622233/00:30
sean-k-mooneydo you know where the doc for how to fix that are i think i have to run a nova-manage command but i cant remember00:31
sean-k-mooneyi think i need to run nova-manage cell_v2 discover_hosts?00:32
dansmithsean-k-mooney: discover-hosts?00:33
sean-k-mooneynova-manage cell_v2 list_cells00:33
dansmithnova-manage cell_v2 discover_hosts --verbose00:33
sean-k-mooneynova-manage cell_v2 discover_hosts --verbose00:33
sean-k-mooneywronge terminal00:33
dansmithheh00:33
dansmither, I mean:00:33
dansmithPassword:00:34
sean-k-mooneyhehe00:34
sean-k-mooneyya that fixed it00:34
dansmithcool00:34
sean-k-mooneyam i guess this might be a bug in devstack then.00:34
dansmithsean-k-mooney: it's just because you have multiple nodes00:35
dansmithdevstack-gate does the final discover after both are done in that case00:35
sean-k-mooneyyes though i would have expected this to be part the n-cpu service install00:35
dansmithwell,00:36
dansmithin our multi-node setups, the second node can't see the database on purpose, so it can't do that00:36
dansmithbut I guess you could make it conditional in devstack so it was graceful00:36
sean-k-mooneyoh ya right so it has to run on the controler00:36
dansmithyeah00:36
sean-k-mooneyok that makes sense well i know for the future havent done multi node with devstack in a while00:37
sean-k-mooneythanks for the help00:37
*** zhurong has joined #openstack-nova00:38
*** itlinux has joined #openstack-nova00:38
*** smatzek has joined #openstack-nova00:38
*** thorst has quit IRC00:40
*** smatzek has quit IRC00:41
*** yingjun has quit IRC00:43
dansmithnp00:46
*** thorst has joined #openstack-nova00:47
*** thorst has quit IRC00:51
*** erlon has quit IRC00:53
*** thorst has joined #openstack-nova00:58
*** thorst has quit IRC01:00
*** thorst has joined #openstack-nova01:00
*** thorst_ has joined #openstack-nova01:01
*** swebster_ has joined #openstack-nova01:01
*** swebster has quit IRC01:02
*** huanxie has joined #openstack-nova01:04
*** thorst has quit IRC01:04
*** crushil has joined #openstack-nova01:04
*** thorst_ has quit IRC01:05
*** TuanLA has joined #openstack-nova01:06
*** phuongnh has joined #openstack-nova01:08
*** mriedem_dinner is now known as mriedem01:11
openstackgerritTakashi NATSUME proposed openstack/nova master: List/show all server migration types (1/2)  https://review.openstack.org/43060801:14
*** coreywright has quit IRC01:14
*** chyka has joined #openstack-nova01:17
*** yangyape_ has joined #openstack-nova01:17
openstackgerritNaichuan Sun proposed openstack/nova master: xenapi: Support live migration in pooled multi-nodes environment  https://review.openstack.org/48945101:18
*** gjayavelu has quit IRC01:19
openstackgerritTakashi NATSUME proposed openstack/nova master: List/show all server migration types (1/2)  https://review.openstack.org/43060801:19
*** gbarros has quit IRC01:22
*** Apoorva has quit IRC01:25
*** litao__ has joined #openstack-nova01:25
*** Shunli has joined #openstack-nova01:27
*** sbezverk has joined #openstack-nova01:31
*** Sukhdev has quit IRC01:31
*** coreywright has joined #openstack-nova01:33
*** zhouyaguo has joined #openstack-nova01:35
*** yamamoto has joined #openstack-nova01:36
*** markvoelker has joined #openstack-nova01:38
*** itlinux has quit IRC01:39
*** yamamoto has quit IRC01:44
*** zhurong has quit IRC01:50
*** zhurong has joined #openstack-nova01:59
*** thorst has joined #openstack-nova02:02
*** gyee has quit IRC02:09
*** hongbin has joined #openstack-nova02:13
*** markvoelker has quit IRC02:13
*** dave-mcc_ has quit IRC02:13
*** ijw has quit IRC02:16
*** ijw has joined #openstack-nova02:16
*** ijw has quit IRC02:17
*** hongbin_ has joined #openstack-nova02:19
*** rcernin has joined #openstack-nova02:20
*** hongbin has quit IRC02:21
*** Tom has joined #openstack-nova02:23
*** itlinux has joined #openstack-nova02:24
*** chyka has quit IRC02:25
*** spotz has quit IRC02:29
*** lbragstad has joined #openstack-nova02:31
*** zhouyaguo has quit IRC02:33
*** zhouyaguo has joined #openstack-nova02:33
*** yamamoto has joined #openstack-nova02:36
*** spotz has joined #openstack-nova02:37
*** mingyu has quit IRC02:44
*** mingyu has joined #openstack-nova02:47
*** esberglu has joined #openstack-nova02:49
*** itlinux has quit IRC02:50
openstackgerritTakashi NATSUME proposed openstack/nova master: List/show all server migration types (2/2)  https://review.openstack.org/45948302:52
*** esberglu has quit IRC02:53
*** armax has quit IRC03:04
*** markvoelker has joined #openstack-nova03:10
openstackgerritTakashi NATSUME proposed openstack/python-novaclient master: Microversion 2.54 - List/Show all server migration types  https://review.openstack.org/43083903:10
*** zhurong has quit IRC03:12
*** baoli has quit IRC03:26
*** hongbin_ has quit IRC03:28
*** markvoelker has quit IRC03:42
openstackgerritPhilip Choi proposed openstack/nova master: Add warning message for cache_utils connection failure  https://review.openstack.org/50192803:47
*** cfriesen has quit IRC03:48
*** udesale has joined #openstack-nova03:48
*** mdnadeem has joined #openstack-nova03:50
*** lbragstad has quit IRC04:00
*** links has joined #openstack-nova04:02
*** zhurong has joined #openstack-nova04:21
*** zhouyaguo has quit IRC04:26
*** zhouyaguo has joined #openstack-nova04:26
*** hieulq has quit IRC04:27
*** trungnv has quit IRC04:27
*** TuanLA has quit IRC04:27
*** trungnv has joined #openstack-nova04:28
*** TuanLA has joined #openstack-nova04:28
*** hieulq has joined #openstack-nova04:28
*** yangyape_ has quit IRC04:33
*** yangyapeng has joined #openstack-nova04:33
*** esberglu has joined #openstack-nova04:37
*** markvoelker has joined #openstack-nova04:39
*** esberglu has quit IRC04:42
*** Sukhdev has joined #openstack-nova04:43
openstackgerritzhangyangyang proposed openstack/nova master: Move libvirts qemu-img support to privsep  https://review.openstack.org/50784804:55
*** psachin has joined #openstack-nova04:56
*** sridharg has joined #openstack-nova05:03
*** markvoelker has quit IRC05:12
*** ratailor has joined #openstack-nova05:14
*** Shunli has quit IRC05:15
*** rcernin has quit IRC05:15
*** zhouyaguo has quit IRC05:18
*** zhouyaguo has joined #openstack-nova05:18
*** armax has joined #openstack-nova05:23
*** pcaruana has joined #openstack-nova05:24
*** stvnoyes has quit IRC05:26
*** yangyapeng has quit IRC05:27
*** yangyapeng has joined #openstack-nova05:27
*** takashin has left #openstack-nova05:27
*** stvnoyes has joined #openstack-nova05:29
*** pcaruana has quit IRC05:29
*** vishwana_ has joined #openstack-nova05:35
*** lpetrut has joined #openstack-nova05:35
*** vishwan__ has joined #openstack-nova05:36
*** zhouyaguo has quit IRC05:36
*** zhouyaguo has joined #openstack-nova05:37
*** vishwanathj has quit IRC05:37
*** Sukhdev has quit IRC05:37
*** vishwana_ has quit IRC05:39
*** yangyapeng has quit IRC05:39
*** yangyapeng has joined #openstack-nova05:41
*** yangyape_ has joined #openstack-nova05:44
*** yangyapeng has quit IRC05:44
*** Tom has quit IRC05:44
*** Tom has joined #openstack-nova05:45
*** moshele has joined #openstack-nova05:51
*** artom has joined #openstack-nova05:52
*** rcernin has joined #openstack-nova05:52
*** artom has quit IRC05:52
*** artom has joined #openstack-nova05:53
*** yamamoto_ has joined #openstack-nova05:53
*** yangyape_ has quit IRC05:53
*** artom has quit IRC05:53
*** artom has joined #openstack-nova05:54
*** artom has quit IRC05:54
*** artom has joined #openstack-nova05:55
*** artom has quit IRC05:55
*** artom has joined #openstack-nova05:56
*** artom has quit IRC05:56
*** yamamoto has quit IRC05:57
*** artom has joined #openstack-nova05:57
*** josecastroleon has quit IRC06:00
*** yangyapeng has joined #openstack-nova06:01
*** yangyape_ has joined #openstack-nova06:02
*** lpetrut has quit IRC06:02
*** sahid has joined #openstack-nova06:04
*** yangyapeng has quit IRC06:06
*** ratailor is now known as rtailor06:06
*** lpetrut has joined #openstack-nova06:07
*** markvoelker has joined #openstack-nova06:09
*** Oku_OS-away is now known as Oku_OS06:14
*** masber has joined #openstack-nova06:15
*** crushil has quit IRC06:19
*** jmlowe has quit IRC06:20
*** esberglu has joined #openstack-nova06:25
*** andreas_s has joined #openstack-nova06:29
*** esberglu has quit IRC06:29
*** mdnadeem has quit IRC06:30
*** lpetrut has quit IRC06:38
*** artom has quit IRC06:42
*** markvoelker has quit IRC06:43
*** artom has joined #openstack-nova06:43
*** artom has quit IRC06:44
*** artom has joined #openstack-nova06:45
*** artom has quit IRC06:45
*** mdnadeem has joined #openstack-nova06:45
*** artom has joined #openstack-nova06:46
*** avolkov has joined #openstack-nova06:48
*** mdnadeem_ has joined #openstack-nova06:48
*** yangyape_ has quit IRC06:51
*** mdnadeem has quit IRC06:51
*** yingjun has joined #openstack-nova06:51
*** yangyapeng has joined #openstack-nova06:51
*** pcaruana has joined #openstack-nova07:04
*** lpetrut has joined #openstack-nova07:11
*** yangyapeng has quit IRC07:13
*** yangyapeng has joined #openstack-nova07:14
*** yangyape_ has joined #openstack-nova07:24
*** lpetrut has quit IRC07:27
*** yangyapeng has quit IRC07:27
*** lpetrut has joined #openstack-nova07:27
*** phuongnh has quit IRC07:31
*** TuanLA has quit IRC07:31
*** masber has quit IRC07:31
*** TuanLA has joined #openstack-nova07:32
*** phuongnh has joined #openstack-nova07:32
*** tssurya has joined #openstack-nova07:32
*** yangyapeng has joined #openstack-nova07:32
*** yangyap__ has joined #openstack-nova07:34
*** yangyapeng has quit IRC07:34
*** yangyape_ has quit IRC07:35
*** lpetrut has quit IRC07:39
*** jpena|off is now known as jpena07:39
*** markvoelker has joined #openstack-nova07:40
*** Guest93 has quit IRC07:42
*** yangyapeng has joined #openstack-nova07:44
*** yangyap__ has quit IRC07:44
*** yangyapeng has quit IRC07:44
*** yangyapeng has joined #openstack-nova07:45
*** yangyape_ has joined #openstack-nova07:49
*** yangyapeng has quit IRC07:49
*** armax has quit IRC07:57
*** mdnadeem_ has quit IRC08:00
*** hferenc has joined #openstack-nova08:01
*** moshele has quit IRC08:03
*** phuongnh has quit IRC08:03
*** TuanLA has quit IRC08:03
*** TuanLA has joined #openstack-nova08:04
*** phuongnh has joined #openstack-nova08:04
*** yangyapeng has joined #openstack-nova08:07
*** yangyape_ has quit IRC08:09
openstackgerritmelanie witt proposed openstack/nova master: WIP Improve the CellDatabases test fixture and usage  https://review.openstack.org/50843208:09
*** markvoelker has quit IRC08:12
*** esberglu has joined #openstack-nova08:13
*** esberglu has quit IRC08:18
*** zhouyaguo has quit IRC08:23
*** ralonsoh has joined #openstack-nova08:25
*** zhouyaguo has joined #openstack-nova08:30
*** yangyapeng has quit IRC08:36
*** yangyapeng has joined #openstack-nova08:36
*** jaosorior has joined #openstack-nova08:37
*** alexchadin has joined #openstack-nova08:37
*** belmoreira has joined #openstack-nova08:38
*** sbezverk has quit IRC08:38
*** trinaths has joined #openstack-nova08:49
*** sapcc-bot has joined #openstack-nova08:50
*** dgonzalez_ has joined #openstack-nova08:50
*** mkoderer_ has joined #openstack-nova08:50
*** carthaca_ has joined #openstack-nova08:50
*** david_1 has joined #openstack-nova08:50
*** carthaca_ has quit IRC08:52
*** mkoderer_ has quit IRC08:52
*** dgonzalez_ has quit IRC08:52
*** david_1 has quit IRC08:52
*** chyka has joined #openstack-nova08:55
*** chyka has quit IRC09:00
*** mriedem has quit IRC09:01
*** yamamoto_ has quit IRC09:01
*** andreas_s_ has joined #openstack-nova09:01
*** andreas_s has quit IRC09:05
*** markvoelker has joined #openstack-nova09:09
*** Tom has quit IRC09:15
*** Tom has joined #openstack-nova09:16
*** Tom has quit IRC09:21
*** yamamoto has joined #openstack-nova09:29
*** mingyu_ has joined #openstack-nova09:29
*** mingyu has quit IRC09:32
*** yamamoto has quit IRC09:32
*** mingyu has joined #openstack-nova09:32
*** mingyu_ has quit IRC09:36
*** mingyu_ has joined #openstack-nova09:38
*** yamamoto has joined #openstack-nova09:40
*** mingyu has quit IRC09:42
*** udesale has quit IRC09:42
*** sambetts|afk is now known as sambetts09:42
*** markvoelker has quit IRC09:43
*** mingyu_ has quit IRC09:45
*** alexchadin has quit IRC09:47
*** alexchadin has joined #openstack-nova09:48
*** cdent has joined #openstack-nova09:48
*** yingjun has quit IRC09:52
*** yamamoto has quit IRC09:54
*** mingyu has joined #openstack-nova09:58
*** esberglu has joined #openstack-nova10:01
*** yamamoto has joined #openstack-nova10:01
*** avolkov has quit IRC10:02
*** avolkov has joined #openstack-nova10:02
*** flanders_ has joined #openstack-nova10:04
*** trungnv has quit IRC10:05
*** esberglu has quit IRC10:06
*** jgriffith has quit IRC10:08
* stephenfin would like dansmith to know that he still has dibs on 'nova/conf/network.py' when nova-net goes. dansmith can have _evvvverything_ else :)10:10
*** jgriffith has joined #openstack-nova10:14
*** derekh has joined #openstack-nova10:22
*** trinaths has left #openstack-nova10:23
openstackgerritStephen Finucane proposed openstack/nova master: setup.cfg: Explicitly set [build_sphinx] builders  https://review.openstack.org/50848310:24
*** hieulq has quit IRC10:24
*** mingyu has quit IRC10:24
*** Tom has joined #openstack-nova10:24
*** TuanLA has quit IRC10:27
*** masber has joined #openstack-nova10:28
*** Tom has quit IRC10:29
*** mingyu has joined #openstack-nova10:29
*** masber has quit IRC10:32
*** moshele has joined #openstack-nova10:32
*** mingyu has quit IRC10:34
*** mingyu has joined #openstack-nova10:34
openstackgerritJohn Garbutt proposed openstack/nova-specs master: Support traits in the Ironic driver  https://review.openstack.org/50705210:35
*** yangyapeng has quit IRC10:36
*** zhurong has quit IRC10:36
*** phuongnh has quit IRC10:37
*** yangyapeng has joined #openstack-nova10:37
*** yangyapeng has quit IRC10:38
*** mingyu has quit IRC10:38
*** mingyu has joined #openstack-nova10:38
*** markvoelker has joined #openstack-nova10:40
*** mingyu has quit IRC10:40
*** alexchadin has quit IRC10:41
*** mingyu has joined #openstack-nova10:42
*** mingyu has quit IRC10:46
*** mingyu has joined #openstack-nova10:46
*** mingyu_ has joined #openstack-nova10:48
*** mingyu has quit IRC10:51
cdentgibi: in the nova meeting last night it was mentioned that you need somewhere to host some burndown stuff. I’ve got some space that could work10:55
gibicdent: hi! thanks for the offer.11:02
*** mingyu_ has quit IRC11:03
gibicdent: the burndown is basically a static html + js and a python script called from cron to generate some data that the js code draws11:03
gibicdent: https://github.com/gibizer/nova-versioned-notification-transformation-burndown11:03
gibiit was run on OpenShift 2.0 but that is retired and 3.0 does not support cron jobs in the free tier any more11:04
openstackgerritStephen Finucane proposed openstack/nova master: conf: Fix indentation of database options  https://review.openstack.org/44309711:04
openstackgerritStephen Finucane proposed openstack/nova master: conf: Resolve TODOs in 'database'  https://review.openstack.org/39369511:04
openstackgerritStephen Finucane proposed openstack/nova master: conf: Remove 'db_driver' config opt  https://review.openstack.org/50848711:04
stephenfinbauzas: Easy +W here (it's a typo fix) https://review.openstack.org/#/c/489028/11:05
* cdent looks at the repo11:05
gibicdent: I'm happy to rework the code to fitt it to your env if needed11:06
cdentit looks pretty straightforward, let me do a quick poke11:08
*** alexchadin has joined #openstack-nova11:10
*** markvoelker has quit IRC11:12
*** andreas_s_ has quit IRC11:12
cdentgibi: http://burndown.peermore.com/nova-notification/11:14
cdenthow often should the cron job run?11:14
gibiuntil now it run hourly11:17
*** sapd_ has quit IRC11:17
*** sapd_ has joined #openstack-nova11:17
*** sapd_ has quit IRC11:17
gibiso if that not a problem for you then let's do it that way11:17
gibithanks for moving it to your server11:17
gibi:)11:17
*** sapd_ has joined #openstack-nova11:18
cdentyeah, hourly is fine with me11:18
gibicdent: I really appreciate your hosting. It saved me at least couple of days of work finding another free hosting and adapting to it.11:19
cdentmy pleasure, it’s easy enough to do, you already have all the necessary bits in the repo, very simply put together11:20
cdentshould update every hour on the hour now11:20
gibicdent: sometimes I have to update the to_be_transformed file in the repo (mostly due to decisions not to transform someting). So I will let you know if you need to pull from github11:20
cdentcool, i’ve got it set up so that a pull is all that’s need11:21
gibiawesome11:21
gibithanks again11:21
*** edmondsw has quit IRC11:21
cdentcould probably even cron that too I suppose11:21
cdentbut for now let’s just do it manually by you telling me11:22
gibisure, that totall works for me11:22
* cdent gets VC funding for burndown as a service11:22
*** adisky has quit IRC11:23
*** sapd_ has quit IRC11:23
*** sapd_ has joined #openstack-nova11:24
*** alexchadin has quit IRC11:24
*** Tom__ has joined #openstack-nova11:25
*** smatzek has joined #openstack-nova11:27
*** Tom__ has quit IRC11:30
*** alexchadin has joined #openstack-nova11:30
*** hferenc has quit IRC11:32
*** zhouyaguo has quit IRC11:35
*** yangyapeng has joined #openstack-nova11:35
*** yangyapeng has quit IRC11:36
*** yangyapeng has joined #openstack-nova11:36
*** sri_ has joined #openstack-nova11:37
sri_hey folks: is there any solution for take instance snapshot if instance created form volume ?11:38
*** mingyu has joined #openstack-nova11:40
*** alexchadin has quit IRC11:42
*** alexchadin has joined #openstack-nova11:43
*** mnestratov has joined #openstack-nova11:43
*** jpena is now known as jpena|lunch11:43
openstackgerritRodolfo Alonso Hernandez proposed openstack/os-vif master: Move 'ips' field from Subnet object to VIF object  https://review.openstack.org/50849811:45
*** huanxie has quit IRC11:47
*** chyka has joined #openstack-nova11:49
*** tylerderosagrund has joined #openstack-nova11:51
*** alexchadin has quit IRC11:52
*** litao__ has quit IRC11:52
*** baoli has joined #openstack-nova11:53
*** armax has joined #openstack-nova11:53
*** chyka has quit IRC11:54
*** mingyu has quit IRC11:57
*** liverpooler has joined #openstack-nova11:58
*** thorst has quit IRC12:00
*** thorst has joined #openstack-nova12:00
*** manasm has joined #openstack-nova12:01
*** sapd__ has joined #openstack-nova12:03
*** sapd__ has quit IRC12:03
*** sapd_ has quit IRC12:03
*** sapd__ has joined #openstack-nova12:04
*** esberglu has joined #openstack-nova12:09
*** markvoelker has joined #openstack-nova12:09
*** baoli has quit IRC12:10
*** alexchadin has joined #openstack-nova12:10
gibicdent: I have a question and a nit in https://review.openstack.org/#/c/48520912:10
*** edmondsw has joined #openstack-nova12:13
*** rtailor has quit IRC12:13
* cdent looks12:16
*** MVenesio has joined #openstack-nova12:20
*** edleafe is now known as figleaf12:21
johnthetubaguyfigleaf: I asked an upgrade question on your spec, I missed it last time :(12:25
*** tylerderosagrund has quit IRC12:25
johnthetubaguyfigleaf: its not a hard question, and I left the answer... but we usually require that to be described in the spec.12:25
*** mingyu has joined #openstack-nova12:27
*** markvoelker has quit IRC12:29
* cdent goes on a coffee journey12:29
*** markvoelker has joined #openstack-nova12:29
*** cdent has quit IRC12:29
*** 07IAA8DSW has joined #openstack-nova12:29
*** 5EXAACMRJ has joined #openstack-nova12:29
*** 07IAA8DSW has quit IRC12:33
*** 5EXAACMRJ has quit IRC12:34
*** jmlowe has joined #openstack-nova12:35
*** hemna_ has joined #openstack-nova12:35
*** hoonetorg has joined #openstack-nova12:35
*** hferenc has joined #openstack-nova12:36
*** tonygunk has joined #openstack-nova12:41
*** manasm has quit IRC12:42
*** pchavva has joined #openstack-nova12:44
johnthetubaguycdent: is that what happens you have *lots* of coffee and everything in the world seems fixable?12:44
openstackgerritElod Illes proposed openstack/nova master: WIP: Transform scheduler.select_destinations notification  https://review.openstack.org/50850612:45
figleafjohnthetubaguy: reading and replying now12:45
*** jpena|lunch is now known as jpena12:47
*** smatzek has quit IRC12:48
*** jaypipes has joined #openstack-nova12:48
*** bnemec has joined #openstack-nova12:50
*** armax has quit IRC12:54
*** smatzek has joined #openstack-nova12:55
*** MVenesio has quit IRC12:55
*** lyan has joined #openstack-nova12:56
*** baoli has joined #openstack-nova12:56
*** mriedem has joined #openstack-nova12:57
*** mgariepy has quit IRC12:59
*** mgariepy has joined #openstack-nova13:01
*** ralonsoh_ has joined #openstack-nova13:02
openstackgerritJohn Garbutt proposed openstack/nova master: WIP: possible extra debug info for ironic problem  https://review.openstack.org/50850813:02
johnthetubaguyjaypipes: been double checking the ironic resource provider, its current the ComputeNode uuid as the uuid and the Ironic Node uuid and the name13:03
*** catintheroof has joined #openstack-nova13:03
*** jaypipes is now known as leakypipes13:04
johnthetubaguyleakypipes: I think that's why things change when ironic does its rebalance, the uuid changes, but the name stays the same13:04
leakypipesjohnthetubaguy: on a calll, gimme a few :)13:04
*** cdent has joined #openstack-nova13:05
*** ralonsoh has quit IRC13:05
johnthetubaguyleakypipes: no problems, catch you in a bit13:05
*** smatzek has quit IRC13:06
*** ralonsoh has joined #openstack-nova13:07
*** ralonsoh_ has quit IRC13:08
mriedemjohnthetubaguy: "its current the ComputeNode uuid as the uuid and the Ironic Node uuid and the name"13:09
mriedemthat's generic from the resource tracker i think,13:09
mriedembut good point13:09
mriedembecause it's just using the hypervisor_hostname for the RP name for the other drivers, which is the same as the hostname13:09
mriedembut for ironic it's the uuid13:09
johnthetubaguyyeah, I think its always hypervisor hostname (which can be ironic uuid)13:09
johnthetubaguyyeah, +113:09
johnthetubaguyso this bug is probably more important than I first thought:13:09
johnthetubaguyhttps://bugs.launchpad.net/nova/+bug/171424813:10
openstackLaunchpad bug 1714248 in OpenStack Compute (nova) "Compute node HA for ironic doesn't work due to the name duplication of Resource Provider " [High,Confirmed]13:10
mriedemwe could put another field in the response from the get_inventory method that includes an optional rp_uuid field or something13:10
mriedemto tell the RT what to use13:10
mriedemalthough,13:10
mriedemthat might break how scheduling works13:10
johnthetubaguybut then we will fail looking up the correct rabbit queue I presume?13:10
mriedembecause the scheduler will lookup a compute node from placement's response by uuid13:10
mriedemthis has nothing to do with mq13:11
mriedemare you thinking the cell mapping stuff?13:11
johnthetubaguywell the mapping from the placement object, to the compute node you build the instance on13:11
cdentleakypipes: are the slides from your scheduler wars presentations on teh internets somewhere (not the video)13:11
mriedemcdent: the slides would also be on the summit video site13:11
cdenthmmm13:12
* cdent tries harder13:12
mriedemor not13:12
mriedemhttps://www.openstack.org/videos/boston-2017/scheduler-wars-a-new-hope13:12
mriedemyou'll have to get the licensed version from the author :)13:12
cdentyeah, that’s where I starte4d13:13
*** lbragstad has joined #openstack-nova13:15
*** smatzek has joined #openstack-nova13:17
*** efried is now known as fried_rice13:21
*** eharney has joined #openstack-nova13:22
*** jamesdenton has quit IRC13:25
*** smatzek has quit IRC13:27
*** smatzek has joined #openstack-nova13:28
johnthetubaguyfried_rice: added more details on that ironic rebalance bug from my findings this morning https://bugs.launchpad.net/nova/+bug/1714248:13:28
openstackLaunchpad bug 1714248 in OpenStack Compute (nova) "Compute node HA for ironic doesn't work due to the name duplication of Resource Provider " [High,Confirmed]13:28
* fried_rice looks...13:28
johnthetubaguyadded a patch with some comments in to make it clear where I believe the error occurs13:29
openstackgerritElod Illes proposed openstack/nova stable/ocata: WIP: Functional test for regression bug #1713783  https://review.openstack.org/50516013:30
openstackbug 1713783 in OpenStack Compute (nova) ocata "After failed evacuation the recovered source compute tries to delete the instance" [High,Triaged] https://launchpad.net/bugs/171378313:30
*** slaweq has quit IRC13:30
*** stephenfin is now known as finucannot13:30
fried_ricejohnthetubaguy Okay, haven't dug in fully, but...13:32
*** burt has joined #openstack-nova13:32
fried_ricejohnthetubaguy It is legal to rename a RP.13:32
fried_riceWould that solve it, or do you actually need to change the RP's UUID too?13:32
johnthetubaguyname is the same, uuid is different13:32
cdentmriedem: in case you missed it before gibi and I got http://burndown.peermore.com/nova-notification/ going13:32
fried_riceboo13:32
*** smatzek has quit IRC13:32
johnthetubaguyyeah :(13:32
cdentusing his little framework would be easy to burn lots of other things down13:32
fried_ricejohnthetubaguy And the original UUID comes from where?13:33
johnthetubaguyfried_rice: ComputeNode.uuid13:33
fried_riceWhich makes sense, I suppose.13:33
johnthetubaguyish, yeah13:33
fried_ricejohnthetubaguy Is there an entity in this world that stays the same on this failover deal?13:34
*** dansmith is now known as superdan13:34
superdanfinucannot: ack, I'll allow you to retain rights to conf/network13:35
johnthetubaguyfried_rice: ironic node uuid, i.e. ComputeNode.hypervisor_hostname and ResourceProvider.name13:35
*** psachin has quit IRC13:36
fried_ricejohnthetubaguy Ohh, so hang on, it really *doesn't* make much sense for the RP UUID to be the compute host UUID.  Or at least, it would make just as much sense for it to be the ironic node UUID.13:37
fried_riceThough tbh, I'm a tad confused as to why those aren't the same thing.13:37
johnthetubaguyfried_rice: except when get the thing back, the uuid, we want it to always be the same thing, in our case a ComputeNode13:37
*** ijw has joined #openstack-nova13:37
johnthetubaguyI am not 100% sure at this point13:38
fried_riceDangit, I've slept since we started this conversation, need to get the model straight in my head again.13:38
fried_riceHardware-wise, there's a set of machines.  These are called "ironic nodes".13:38
johnthetubaguyoh wait, that could be the answer, for ironic, we could mess with the ComputeNode.uuid13:38
fried_riceEach "ironic node" gets modeled as a separate RP, right?13:39
johnthetubaguyfried_rice: https://developer.openstack.org/api-ref/baremetal/#list-nodes13:39
johnthetubaguyyeah13:39
*** links has quit IRC13:39
mriedemcdent: nice13:39
fried_riceAnd the nova-compute process runs... where?  On a totally separate machine?13:39
cdentmriedem: runs on a cron job every hour. took an _entire five minutes_ to set up13:40
* cdent is exhausted13:40
mriedemtake a break13:40
*** kenperkins has joined #openstack-nova13:40
johnthetubaguyfried_rice: yeah, noramlly on some controller node13:41
fried_ricejohnthetubaguy Which is called ComputeNode?13:41
johnthetubaguynot really13:42
johnthetubaguyComputeNode is the Nova DB table13:42
johnthetubaguythere is one entry for each ironic node13:42
openstackgerritEd Leafe proposed openstack/nova-specs master: Return Alternate Hosts  https://review.openstack.org/50427513:42
johnthetubaguyi.e. one n-cpu Service has one or more compute nodes13:43
fried_ricejohnthetubaguy So why isn't the ComputeNode.uuid the same as the ironic node UUID, if they represent the same thing?13:43
johnthetubaguyfried_rice: I think just history, I am just looking at how hard that would be right now13:43
superdanbecause computenode is a nova structure13:44
superdanwhich is used for everything else too,13:44
superdanand because you could have generated those uuids from the existing records when things were offline during a save or forced migration13:44
*** gouthamr has joined #openstack-nova13:44
johnthetubaguysuperdan: agreed13:45
openstackgerritMatt Riedemann proposed openstack/nova master: Remove old compat code from servers ViewBuilder._get_metadata  https://review.openstack.org/50832613:45
openstackgerritMatt Riedemann proposed openstack/nova master: Stop joining on system_metadata when listing instances  https://review.openstack.org/50833513:45
openstackgerritMatt Riedemann proposed openstack/nova master: Remove system_metadata loading in Instance._load_flavor  https://review.openstack.org/50835713:45
johnthetubaguysuperdan: it could be the simplest fix, but all the fixes seem terrible13:46
superdanI wasn't following what the problem was13:46
johnthetubaguyhttps://bugs.launchpad.net/nova/+bug/171424813:46
openstackLaunchpad bug 1714248 in OpenStack Compute (nova) "Compute node HA for ironic doesn't work due to the name duplication of Resource Provider " [High,Confirmed]13:46
mriedemthe rp uuid isn't the ironic node uuid13:46
mriedemit's the compute node uuid13:46
johnthetubaguyactually, that fix isn't simple in the upgrade case, naturally13:46
mriedemjohnthetubaguy: was going to say,13:46
mriedemexisting computes...13:46
mriedemyou could change the compute node uuid...but not sure what weirdness would happen from that13:47
superdanthat's hard13:47
superdanI don't think changing the uuid of an existing provider, especially with allocations makes sense13:47
*** gbarros has joined #openstack-nova13:47
*** Dinesh_Bhor has quit IRC13:48
superdanthe ironic driver probably needs to figure this out some way to avoid having to re-do the allocations13:48
johnthetubaguyproblem is the allocation is on the ComputeNode.uuid13:48
superdanotherwise we have a race I think13:48
johnthetubaguyyeah, the race is horrible13:48
superdanright I know13:48
* johnthetubaguy just realises the ironic driver hashring feature means you can't aggregate a multiple ironic deploys in a single region any more, boo13:50
superdanI'm not sure what that means13:50
superdanso one thing we could do,13:50
leakypipesjohnthetubaguy: nova host aggs never worked with ironic nodes, though.13:51
superdanis make sure that the RP name is the uuid of the ironic node13:51
leakypipesjohnthetubaguy: just got off the call, btw, reading back13:51
superdanwhich would be a path for the ironic driver getting the node to find out the uuid of the matching provider13:51
fried_riceCan we do a fake migration, using the "migration_uuid" thingy, to avoid the race?13:51
johnthetubaguythat is true today13:51
superdanfried_rice: doesn't really help13:51
mriedemright the ironic rp name is the ironic node uuid today13:52
johnthetubaguyoh, you mean when we create the new compute node, we use the existing uuid, so we just rename the compute node?13:52
mriedemsince the rp.name == compute_node.hypervisor_hostname == ironic node.uuid13:52
superdanmriedem: is it?13:52
mriedemyar13:52
johnthetubaguyyeah, thats correct13:52
superdanwell in that case, we can just look it up and reparent the compute node object then13:52
superdancomputenode.host = self.host13:52
superdandone13:52
johnthetubaguyyeah, instead of add a new one13:52
johnthetubaguythat's what I just attempted to say, badly13:53
superdanI was going to suggest that we set it so we could do that once it was, but if it's already done we should be able to do that right away13:53
johnthetubaguyI will take a look at doing that13:54
johnthetubaguyin the ironic driver, or where we create the compute nodes already?13:54
superdanum13:54
*** awaugama has joined #openstack-nova13:54
superdanI'd have to go look at that.. I guess there's a bit of a layering violation in there some where13:54
superdanmaybe we can just make compute manager check for this situation and do the re-homing13:55
*** cfriesen has joined #openstack-nova13:56
johnthetubaguyyeah, I just worry about hostname clashes13:56
johnthetubaguyin none ironic cases13:56
*** takedakn has joined #openstack-nova13:56
superdanwell,13:56
superdanthe hostnames have to be unique today or nothing works anyway13:56
superdanbut you could also is_uuid_like() on it for a total hack :)13:57
johnthetubaguygood point, placement has that constraint now, hence the bug in the first place13:57
superdanor set some capability attribute on the driver: CAN_REBALANCE_NODES13:57
superdanLIKES_TO_SHARE=True13:57
johnthetubaguyheh13:57
superdanPLAYS_WELL_WITH_OTHERS=True13:57
johnthetubaguythat doesn't sound right...13:57
*** hongbin has joined #openstack-nova13:57
superdanwe have to have unique hostnames for rpc to work anyway13:58
superdanmaybe not hypervisor_hostname I guess, but they should be the same for non-ironic ones anyway right?13:58
johnthetubaguyisn't that one on the service, but yeah thats the same for regular virt13:58
johnthetubaguyOK, I am convinced, will give that a whirl13:58
*** sbezverk has joined #openstack-nova13:59
superdanyes it's service.host that has to be unique but should be the same for virt systems13:59
*** smatzek has joined #openstack-nova14:00
johnthetubaguycool, just making sure, keep getting it all twisted14:00
*** ralonsoh_ has joined #openstack-nova14:02
*** crushil has joined #openstack-nova14:03
cdentregister headline: OpenStack switches back to twisted14:03
*** takedakn has quit IRC14:03
superdanmriedem: I hadn't seen your second flavor sysmeta patch.. I didn't test with it applied as well14:06
*** ralonsoh has quit IRC14:06
*** sapcc-bot has quit IRC14:06
johnthetubaguycdent: heh14:06
*** sapcc-bot has joined #openstack-nova14:06
mriedemsuperdan: that shouldn't make a difference during instance list14:07
*** dgonzalez_ has joined #openstack-nova14:07
*** tpatzig_ has joined #openstack-nova14:07
*** mkoderer_ has joined #openstack-nova14:07
*** david_ has joined #openstack-nova14:07
*** seife_ has joined #openstack-nova14:07
*** carthaca_ has joined #openstack-nova14:07
mriedemsince it's just used during (1) save and (2) lazy-loading of flavor i think14:07
superdanmriedem: well, unless we started lazy-loading it, but yeah, I was checking for that in the logs anyway14:07
*** dgonzalez_ has quit IRC14:09
*** tpatzig_ has quit IRC14:09
*** mkoderer_ has quit IRC14:09
*** david_ has quit IRC14:09
*** seife_ has quit IRC14:09
*** carthaca_ has quit IRC14:09
*** alexchadin has quit IRC14:09
figleafcdent: heh, I remember the discussions that led to the switch from twisted to eventlet14:11
openstackgerritMerged openstack/nova-specs master: Update a URL  https://review.openstack.org/48902814:13
*** hemna_ has quit IRC14:15
*** mriedem1 has joined #openstack-nova14:20
*** tonygunk has quit IRC14:21
*** tonygunk has joined #openstack-nova14:21
*** kfarr has joined #openstack-nova14:22
*** mriedem has quit IRC14:23
bauzassuperdan: https://review.openstack.org/#/c/498948/10/nova/compute/manager.py@3640 I'm not super expert of all the migration states but I guess a migration revert is not having an 'in-progress state' ?14:25
bauzasI just want to make sure that we don't do the math if we're at the middle of a revert14:25
bauzaswait14:26
bauzasif we get allocations, we know we have nothing to do14:26
*** Yikun__ has joined #openstack-nova14:26
bauzasso we shouldn't care of the migration state then ?14:26
*** moshele has quit IRC14:29
superdanbauzas: right that's just a short-circuit14:29
*** takedakn has joined #openstack-nova14:31
*** flanders_ has quit IRC14:32
*** jaosorior has quit IRC14:33
*** takedakn has quit IRC14:34
*** mriedem1 is now known as mriedem14:43
openstackgerritpriyaduggirala proposed openstack/nova master: Rename arguments in call() of nova/image/glance.py  https://review.openstack.org/50853314:48
*** chyka has joined #openstack-nova14:48
bauzassuperdan: corrolary question, if we are reverting and if we have allocations related to the migration, we of course don't remove them. But then, we will remove them once it's calling finish_revert_resize() right?14:50
leakypipescdent: I presume you found the scheduler presentation links? the links to the presentations are http://bit.ly/scheduler-wars-a-new-hope and http://bit.ly/scheduler-wars-revenge-of-the-split14:50
bauzassuperdan: nevermind, just saw https://review.openstack.org/#/c/498949/11/nova/compute/manager.py14:50
bauzasokay, so +W14:50
cdentleakypipes: i did not, thank you14:50
superdanbauzas: check out the functional tests.. unless you find a gap, they're pretty obsessive about the allocations at all the steps14:51
bauzascool with me14:51
leakypipescdent: BTW, thanks again for continuing your resource provider update emails. they really are very helpful.14:51
leakypipescdent: also, nice use of the word "passel"14:52
cdentleakypipes: you’re welcome. I decided this week I’d go short and focused, in part because that seemed like good timing, but also because sometimes I’m so far behind myself that I haven’t got time to do the big version. we are moving a _ton_ of code these days14:52
*** tonygunk has quit IRC14:56
*** tonygunk has joined #openstack-nova14:57
*** Tom___ has joined #openstack-nova14:58
melwittsuperdan: I dunno if you saw this but I think I found a bug with the get_instance_object_sorted when there are faults. I explained it here https://review.openstack.org/#/c/505417/7/nova/tests/functional/compute/test_instance_list.py@45314:58
superdanmelwitt: so we're not going to merge the patch that pre-joins faults14:59
superdanso that test can go away14:59
superdanis there actually a problem based on how the API works now, or just that that test failed?15:00
melwittsuperdan: it's that _from_db_object doesn't accept pre-joined attrs, i.e. it won't set them. so even after you pre-join it lazy-loads it. and with an untargeted _context, it won't be able to and will get a None fault15:01
*** rcernin has quit IRC15:01
superdanand that test was intentionally not 3 cells (see mriedem's comment above) but only because it wasn't really needed15:01
superdanmelwitt: okay but the api isn't asking for fault to be joined15:01
*** baoli has quit IRC15:01
superdanmelwitt: the api loads the faults it wants in a batch after having collected the full list15:02
melwittsuperdan: you mean nova-api?15:02
superdancompute/api and above but yeah15:02
melwittokay15:02
*** baoli has joined #openstack-nova15:02
superdanso you can delete that test if you want and I can re-add it if I make the fault-joining patch work better15:02
*** baoli has quit IRC15:02
*** Tom___ has quit IRC15:02
superdanmelwitt: this is out of the stack at the moment: https://review.openstack.org/#/c/505456/1015:03
*** tonygunk has quit IRC15:03
melwittsuperdan: okay, so you're saying get_instance_object_sorted won't ever be used the way it is in that test15:03
superdannot until I fix the above patch yeah15:04
melwittokay. yeah, I guess the only thing to look out for is if you want lazy-loads to work on the resulting cross-cell list, you'll have to make sure each of their _context are targeted to their cell15:05
superdanyeah I'm not sure why they wouldn't be in this case, but it's not a scenario we request at the moment anyway15:05
melwittI think each Instance object is inited with the same untargeted context15:06
superdanyou could make the test use 3 cells and make it an xfail with a comment and I can come back to it when I work on the patch if you want15:06
melwittso lazy-loads wouldn't work15:06
superdanmelwitt: but it should be initialized with the context that scatter_gather_all_cells gave to the ...15:06
superdanohhh15:06
superdanI see15:06
superdanyep, I get it15:06
melwittk15:06
superdananyway, feel free to disable or remove that test for your immediate purposes and I'll circle back to it when I try to optimize the fault bit15:07
openstackgerritChris Dent proposed openstack/nova master: [placement] gabbi tests for shared custom resource class  https://review.openstack.org/48520915:07
*** baoli has joined #openstack-nova15:08
*** hemna_ has joined #openstack-nova15:08
mriedemfor anyone that cares, the change to get this legacy nnet job out of nova is here https://review.openstack.org/#/c/508519/15:09
mriedemwe're blocked until then15:09
*** coreywright has quit IRC15:10
superdanbummer15:10
*** Oku_OS is now known as Oku_OS-away15:11
mriedemsuperdan: btw, i was going to propose a forum session to brainstorm ideas for some automated perf testing like what was done this week, for a few reasons,15:11
mriedemlike automating it, but also find out what other projects have done, if anything,15:11
superdanmriedem: we've discussed that a few times, and in boston even15:12
melwittsome other projects use rally for that15:12
superdanalways comes back to not having stable enough stuff15:12
mriedemand because there are several work group sessions, like the public cloud one, where they are talking about "what features do we want?!" and i want to say "what scale testing do we need?!"15:12
mriedemthis wouldn't be voting15:12
superdanit's just a matter of having useful data, and as you saw, it's hard to compare two runs15:13
superdaneven if not voting, if you can't actually draw conclusions...15:13
mriedemwas thinking experimental queue so it's on-demand for things we know might impact performance15:13
superdanyou can compare normalized metrics like number of db queries or something, but runtime and cpu usage are not really doable without dedicated hardware15:13
mriedemmy thought was the job pulls master, runs some baselines, then applies the change, runs the same tests and compares for the relative difference15:13
superdannoisy neighbor problems will still skew those15:14
*** belmoreira has quit IRC15:14
superdanthe window is smaller, granted, but..15:14
mriedemok, so maybe a requirement is it runs on baremetal?15:14
superdanthat would be better yeah15:14
mriedempoint being,15:14
mriedemthere is an obvious need,15:14
superdanthey could be done on virt, but only if salt is applied to the result15:14
mriedemso if we had requirements to start, we could maybe do something15:14
mriedemi'm thinking super simple to start15:15
superdanare you saying my box can't satisfy all of nova's perf testing needs?15:15
mriedemsure, if you want to become the request inbox for every time we need that15:15
superdanheh15:15
*** penick has joined #openstack-nova15:16
mriedemmelwitt: re the rally job, i'd be interested to know if (1) it still works and (2) if so, what does it do? like what benchmark does it compare against?15:16
melwittmriedem: fwiw, I was thinking the same thing recently. if we had some basic perf tests that timed instance list, delete, boot. the main things15:16
mriedemandreykurilin: ^ maybe you can answer that15:16
*** vvargaszte has joined #openstack-nova15:17
*** yamamoto has quit IRC15:18
andreykurilinmriedem: hi! what is the question? :)15:18
andreykurilinmelwitt, mrieden: rally has a bunch of scenarios related to nova15:19
melwittmriedem: I don't know much about it other than I know some projects use it to have some monitor of perf. lemme see if I can find something real quick15:19
mriedemandreykurilin: "re the rally job, i'd be interested to know if (1) it still works and (2) if so, what does it do? like what benchmark does it compare against?"15:20
andreykurilinmriedem: so rally job can be used to check that things are working under some load and in concurrency mode. Also, you can specify SLAs for a single actions, like neutron job did https://github.com/openstack/neutron/blob/master/rally-jobs/neutron-neutron.yaml#L20-L2115:21
melwittmriedem: found one with cinder, example patch https://review.openstack.org/#/c/501478 and the job http://logs.openstack.org/78/501478/1/check/gate-rally-dsvm-cinder-ubuntu-xenial-nv/e0eb0eb/15:21
*** mingyu has quit IRC15:21
*** dave-mccowan has joined #openstack-nova15:22
*** mingyu has joined #openstack-nova15:22
andreykurilinmelwitt: neutron has voting rally job for a year I think15:22
melwittokay, so neutron and cinder were the projects I was probably thinking of15:22
andreykurilincinder has non-voting job for years15:22
andreykurilinmanila has the rally job too15:23
*** coreywright has joined #openstack-nova15:23
mriedemgibi: is this a co-worker of yours? http://forumtopics.openstack.org/cfp/details/5015:23
mriedemandreykurilin: that limit has to be pretty high doesn't it if the rally jobs are running on vms with probably wild variance15:24
mriedemthat was always the reason we didn't have a voting rally job in nova15:24
mriedemand if the limit is really high, then the only things your catching are something that's really crazy and has gone off the rails....15:25
andreykurilinmriedem: `that was always the reason we didn't have a voting rally job in nova` it is wrong. There were several guys who dislike rally from the beggining (I do not want to name them) and blocked everything related15:26
mriedemcdent: are you one of the people that are responsible for filtering these forum session proposals?15:26
cdentno sir15:26
mriedemandreykurilin: ok, personal vendettas aside, i think the general reason was the variance between runs15:27
cdentbut I was reading them anyway, so I thought I’d express an opinion, because why not?15:27
sean-k-mooneyleakypipes: sahid: sahid i do not want to top post over latest responce to vgpu implementation but can your responded to  http://lists.openstack.org/pipermail/openstack-dev/2017-September/122702.html where i pointed out that any vgpu support we introduce cannot rely on mdev as amd do not use mdevs they use sriov15:28
bauzasoh man, I now need to learn the Zuul language15:28
mriedemmdev or bust15:28
sean-k-mooneyleakypipes: sahid we can use mdev but we have to also support sriov15:29
bauzasmriedem: man, you don't imagine15:29
leakypipessean-k-mooney: I don't think anyone's saying we would *only* support mdev.15:29
andreykurilinmriedem: about metrics and their processing. The SLA is a pluggable thing and you can use it in different ways. (90% percentiles, median...). But I agree that the results can look random due to the hardware, but anyway it can show bad trends, like it was in June - http://andreykurilin.me/trends/trends_gate-rally-dsvm-neutron-rally-ubuntu-xenial.html#/NeutronNetworks.list_agents15:30
leakypipessean-k-mooney: I think everyone's in agreement that we should enable multiple device management APIs, not tie us to one or another.15:30
andreykurilinmriedem: also, the reason of making rally job voting in neutron was an ability to check the concurrency issues15:30
sean-k-mooneyleakypipes: yes15:31
sahidsean-k-mooney: yes really, since the beginning the point was that, vgpus can be exposed with sriov or mdev15:31
cdentsean-k-mooney: while you’re around, can you have a look at https://review.openstack.org/#/c/504540/ you were one of the people who had some ideas on how to do allocation candidate limiting, and there now lots of ideas on that spec, but some confusion on what things we are trying to optimize15:31
sean-k-mooneycdent: sure opening it now15:31
cdentthanks15:32
bauzasleakypipes: sean-k-mooney: that's one of the reasons why the vGPU spec is saying we will only pass GPU resource classes15:32
bauzasif libvirt uses mdev, then meh15:32
bauzasif xen is using anything else, then meh15:32
sean-k-mooneybauzas: its nto a libvirt vs xen thing15:32
bauzaswe shouldn't just leak out the technical details15:32
sahidit's not libvirt who is using mdev, it's the hardware driver...15:32
sean-k-mooneyat the hardware level amd uses hardware based partitioning of the gpu via sriov15:33
bauzaswhatever the solution is, the interface with the compute manager is just a resource class15:33
sean-k-mooneyintel and nvidia do it via the diriver with mdevs15:33
bauzasfor the moment, libvirt supports virtual GPUs by mdevs but that's just a temporary thing AFAICU15:33
andreykurilinmriedem: I agree that it is difficult to compare the results from different hardware(which we have in infra), but it is not a good time for openstack and I do not know the company who are ready to give reserved hardware for a performance ci15:33
bauzassean-k-mooney: yeah that's the driver which provides mdev types, libvirt just gets them15:34
sean-k-mooneybauzas: the hypervisror did nto need modifcation to work with vgpus with sriov15:34
bauzassean-k-mooney: not sure I get your point :)15:34
sean-k-mooneymy point is that libvirt works with amd vgpus also15:35
sean-k-mooneyyou just to a pci passthough of the sriov vf15:35
bauzasah okay, and amd driver uses SR-IOV for that?15:35
bauzasthe AMD driver isn't using mdevs ?15:35
sean-k-mooneyyes they do the virtualisation in hardware15:36
bauzasI see15:36
sean-k-mooneythey do not use mdevs15:36
bauzasso, the thing is15:36
bauzasI see baby steps here15:36
bauzasstep 1/ make sure we have something workable without adding more tech deby15:36
bauzasstep 2/ work on the generic thingy from efried and "do the work" (c)15:37
*** cdent has quit IRC15:37
bauzasstep 3/ migrate how we track PCI devices from the old world to the new world15:37
fried_rice++15:38
bauzashah, Friday!15:38
bauzasholy shit, I said many times I need a ZNC bot for me15:38
*** bauzas is now known as bauwser15:38
mriedemandreykurilin: http://forumtopics.openstack.org/cfp/details/5515:38
mriedemsuperdan: melwitt: ^15:38
melwittcool15:39
sean-k-mooneybauwser: makes sense, my concern is coming for interop side in that i would hope at the api level we will not need to expose if it is an mdev or sriov based vgpu15:40
bauwsersean-k-mooney: that's where I'm pretty firm in my mind15:40
bauwsersean-k-mooney: outside the virt driver, we shouldn't leak out the details15:41
sean-k-mooneybauwser: we may need to expose a vender in some form for guest driver reasons but other then that how we virtualise should be internal to nova15:41
bauwserthe virt driver will just report inventories15:41
bauwserhow those inventories will be populated will be different based on the driver15:41
bauwserbut that will be consistent15:41
openstackgerritEric Fried proposed openstack/nova master: WIP: Use ksa adapter for cinder client  https://review.openstack.org/50834515:42
sean-k-mooneysure and i think we can use traits on the provider to hanel the "i only have a nvida driver in my vm image" problem15:42
fried_ricemordred ^ would you mind checking that I did my version discovery sanely here?15:42
bauwserso that a flavor asking for an amount of 2 of the VGPU resource class and a trait of 'K800' will get 2 K800 vGPUs15:43
bauwserthat's where the API is15:43
bauwseractually a good call for trying to fit the fried_rice's model with vGPUs15:43
*** burt has quit IRC15:44
sean-k-mooneybauwser: perhaps yes or as image metadata where i say i can support x,y,z and the and flavor say i want y and we compare to make sure its compatible15:44
bauwsersean-k-mooney: mmm15:45
bauwsersean-k-mooney: I wonder if that wouldn't be just a filter for the image metadata15:45
bauwserlike, I can find all the hosts that can support y (if y is asked by the flavor)15:45
sean-k-mooneyim thinking we can add traits to the image metadata15:45
bauwserthat is a placement thing15:45
bauwserwell15:46
bauwserhonestly that's a thought15:46
*** ctrath has joined #openstack-nova15:46
*** tonygunk has joined #openstack-nova15:46
mordredfried_rice: looks good on quick glance - but I haven't ooked deeply yet15:47
fried_ricemordred Cool, thanks.15:47
bauwserwell, I need to call it a day, actually15:48
mriedemfried_rice: let's let mordred deal with the infra fire that's raging15:48
mriedem:)15:48
fried_ricemriedem mordred Sorry, didn't smell the smoke.15:49
mriedem#openstack-infra15:49
sean-k-mooneyleakypipes: am qq. is there a way we can retrive the compute node resocue provide by the hostid? and if not would you be against adding one. it could just be a trait15:49
leakypipessean-k-mooney: each resource provider has a name attribute.15:50
leakypipessean-k-mooney: that is set to hypervisor_hostname value15:50
leakypipessean-k-mooney: so yeah, you can already do that.15:50
leakypipessean-k-mooney: and that name is unique BTW15:50
sean-k-mooneyleakypipes: yes i know we require that all hostnames be unique in the openstack cloud15:51
sean-k-mooneyleakypipes: the issue is the get resouce provider api to day i think need you to pass the uuid15:51
mriedemsean-k-mooney: you can find the compute node by hostname, and from the compute node uuid look up the RP15:51
mriedemthat's what our functional tests do15:51
sean-k-mooneyi dont know if we have a rest api call to get it by name15:51
* fried_rice confirms: GET /resource_providers?name={hostname} works15:52
leakypipessean-k-mooney: no, I mean when resource providers are added by the compute node, the resource provider's name attribute is set to the hypervisor_hostname15:52
sean-k-mooneyfried_rice: cool if GET /resource_providers?name={hostname}  works that perferct15:52
leakypipessean-k-mooney: GET /resource_providers?name=$hypervisor_hostname15:52
*** derekh has quit IRC15:52
leakypipessean-k-mooney: sorry, yeah, fried_rice beat me to it.15:52
fried_riceYeah, to be technically correct, in case $hypervisor_hostname isn't the same as `hostname`15:52
leakypipeshe's a wily coyote, that fried_rice15:52
leakypipesfried_rice: which is the case for Ironic nodes :)15:53
fried_riceright15:53
sean-k-mooneythe reason i was asking is for the neutron placement work. when nested provider are added we will need to find the computenode provider for the current host hence question15:53
fried_riceSo as long as you know how virt is reporting the hypervisor hostname, you're golden.15:54
sean-k-mooneyin neutron i belive we have the host id in a config that must be set to match the one used by nova15:54
leakypipesright15:55
fried_rice# Hostname to be used by the Neutron server, agents and services running on15:55
fried_rice# this machine. All the agents and services running on this machine must use15:55
fried_rice# the same host value. (unknown value)15:55
fried_rice#host = example.domain15:55
fried_ricein neutron.conf ?15:55
leakypipesdevice_id == hostname, right?15:55
leakypipessean-k-mooney: ^\15:55
*** Tom___ has joined #openstack-nova15:55
sean-k-mooneydevice_id?15:55
*** ctrath has left #openstack-nova15:56
leakypipessean-k-mooney: in neutron port response...15:56
leakypipessean-k-mooney: never mind.15:57
sean-k-mooneyoh  binding:host_id  is15:57
mriedemsuperdan: http://forumtopics.openstack.org/cfp/details/6015:57
mriedemmelwitt: tssurya: ^15:57
*** zzzeek has quit IRC15:58
* tssurya looking15:58
superdanmriedem: hmm, did the gate just get dumped?15:59
mriedemidk15:59
sean-k-mooneybinding:host_id is set by nova when the port is bound by the compute node15:59
sean-k-mooneyleakypipes: but thanks that answers my question ill let rodolfo know15:59
*** zzzeek has joined #openstack-nova15:59
sean-k-mooneyand thank fried_rice too16:00
melwittI thought I saw something in #openstack-infra that said they were gonna stop zuul v316:00
superdans'empty16:00
*** Tom___ has quit IRC16:00
melwittbut I wasn't sure if I read it right16:00
*** xyang1 has joined #openstack-nova16:00
fried_riceI just got a weird merge conflict16:00
superdanwell, I had to fix the unit test patch anyway, so I'll push up some fresh in a minute16:00
*** huanxie has joined #openstack-nova16:01
openstackgerritDan Smith proposed openstack/nova master: Move cell0 marker test to Cellsv1DeprecatedTestMixIn  https://review.openstack.org/50831416:02
openstackgerritDan Smith proposed openstack/nova master: Use improved instance_list module in compute API  https://review.openstack.org/50541816:02
openstackgerritDan Smith proposed openstack/nova master: Fix CellDatabases fixture swallowing exceptions  https://review.openstack.org/50631216:02
openstackgerritDan Smith proposed openstack/nova master: Fix minor input items from previous patches  https://review.openstack.org/50641616:02
*** smatzek has quit IRC16:03
superdanmriedem: did you want me to do something with those session proposals or just headupping me?16:03
*** smatzek has joined #openstack-nova16:03
mriedemjust fyi16:03
superdanack16:03
*** sbezverk has quit IRC16:04
*** sambetts is now known as sambetts|afk16:05
melwittmriedem: are you going to propose an install/upgrade one or shall I? you didn't mention one in your ML post but I wanted to double check16:05
*** smatzek has quit IRC16:06
*** smatzek has joined #openstack-nova16:06
*** mingyu has quit IRC16:08
*** smatzek has quit IRC16:08
*** smatzek has joined #openstack-nova16:09
*** owalsh_ has joined #openstack-nova16:09
mriedemmelwitt: i wasn't planning on proposing a specific nova upgrades one no, i figured some of that will be covered in cells/placement sessions on what happened in pike and the focus for queens16:09
mriedemfeel free though16:09
openstackgerritJohn Garbutt proposed openstack/nova master: Re-use existing ComputeNode on ironic rebalance  https://review.openstack.org/50855516:10
superdanso,16:10
superdansince we don't get to prioritize these (right?) I'd hate to lose one of these important sessions for an upgrade one that we've hit many times before16:10
superdanif we have space, it's always good to have one of those,16:10
superdanbut if the TC is picking, I don't want them to choose that over, say, the cells or placement one16:10
*** mingyu has joined #openstack-nova16:11
superdancould we maybe indicate a preferred ordering somehow? comments?16:11
*** owalsh has quit IRC16:11
mriedemcomments are about all you get16:11
mriedemNOVA PRIORITY: 516:11
johnthetubaguydid you make the placement one sound cross project-ey?16:11
johnthetubaguythat might be a good way to keep a good session16:11
mriedemi didn't propose a placement one because of http://forumtopics.openstack.org/cfp/details/5016:11
mriedem^ is both generic and too specific16:12
mriedemwhat i think that is, is talking about how nested resource providers and custom resource classes will let you do all kinds of external scheduling things16:12
johnthetubaguyyeah, it is.16:12
mriedemwhich is not what i was going to propose16:12
mriedemi guess i'll just propose my own16:13
johnthetubaguyI would tempted to add a good one, and note its a duplicate?16:13
melwittokay. my thinking was if ppl are recently having trouble installing or upgrading nova, that was important and I wanted to gather action items on whether we need to improve docs, relnotes, nova-manage commands, etc16:13
*** owalsh_ has quit IRC16:14
mriedemmelwitt: bauwser has one for docs and relnotes specifically16:14
mriedemi see your point16:14
mriedemare people recently having trouble beyond ocata?16:15
mriedemi know ocata was tough for some people16:15
melwittyeah, I saw that. docs and relnotes generic session without a target like install/upgrade I'm not sure what will come out of that16:15
mriedem"the release notes don't have enough information!"16:16
mriedem"the release notes have too much information"16:16
mriedemthat's what will come out of that16:16
superdanyep16:16
superdanand what will come of the upgrade one is much of the same we've heard before I think, which is why I care less about it than the other things16:17
melwittbut I understand everyone's point. I really wish I knew if operators were suffering on newton/ocata/pike and what we could do to ease that. I don't remember discussing that span of releases in a session but I do remember a lot of ppl coming into our channel with issues16:17
superdanif we have space for it, then I'm certainly happy to have it16:17
*** vvargaszte has quit IRC16:17
*** yamamoto has joined #openstack-nova16:18
melwittwe responded to the issues and improved our docs and nova-manage commands. and now it's our best assumption that everything is okay but I wasn't sure if it is since we haven't had anyone tell us. it's more like lack of complaint16:18
*** owalsh has joined #openstack-nova16:19
melwittjust thinking aloud mostly16:19
melwitthow can we tell if we have space for more proposals? we can't right?16:20
*** jmlowe has quit IRC16:20
mriedemmelwitt: we can't16:20
mriedemthere is a fast forward upgrades session16:21
mriedemso my guess would be a nova-specific upgrade session would be asking for FF upgrades16:21
mriedemttx: how many slots are there for forum topics?16:21
mriedems/topics/sessions/16:21
mriedemi would also love feedback on where we've sucked it up perf-wise,16:22
mriedembut,16:22
mriedem...16:23
*** itlinux has joined #openstack-nova16:23
melwittyeah. this is weird that we can't control what we have sessions about and have to worry that some sessions would bump others out16:23
*** yamamoto has quit IRC16:24
mriedemand as a result we don't propose as many16:24
mriedemmeanwhile there are like 4+ overlapping sessions about HPC and NFV16:24
itlinuxhello all.. I have a docker image that I installed the clients on and I have the endpoints running on HTTPS.. when I do nova list on that box I get an error.. if I use my virtualbox with openstack installed I can query my remote openstack with the nova list just fine.. any tips?16:25
openstackgerritDan Smith proposed openstack/nova master: Fix minor input items from previous patches  https://review.openstack.org/50641616:25
superdanmriedem: so, in my fixing of your fields thing in ^16:25
superdanI found another place we were implicitly querying for a ton of extra stuff when all we want was metadata16:26
*** pcaruana has quit IRC16:26
*** shewless has joined #openstack-nova16:26
johnthetubaguyfried_rice: leakypipes: superdan: thanks for your help with the ironic placement bug, fix up for review, not had chance to test it yet, something for next week: https://review.openstack.org/#/c/50855516:26
superdan"ton" being "four" instead of "one"16:26
superdanyou know, apply my normal exaggeration descaling filter16:26
superdanjohnthetubaguy: cool16:28
ttxmriedem: not sure, maybe thingee or flaper87 know16:29
*** gbarros has quit IRC16:29
* ttx is not on the organization group this time16:29
mriedemjohnthetubaguy: superdan: leakypipes: cdent: figleaf: fyi http://forumtopics.openstack.org/cfp/details/6016:30
superdanmriedem: you posted that on the cells one, did you mean to do it on a placement one?16:31
*** huanxie has quit IRC16:31
leakypipesack, will do johnthetubaguy16:31
mriedemsuperdan: you mean this? https://review.openstack.org/#/c/506416/13/nova/compute/api.py@375516:32
mriedemcrap yes16:32
superdanmriedem: yes that16:32
mriedemsuperdan: have you seen https://review.openstack.org/#/c/508299/16:33
superdanmriedem: there's a test that failed if we don't pass something to that now that we're doing the fields calculation early, but it means we stop taking the obvious default16:33
superdanmriedem: heh, okay, well, I need it for the change in that patch16:33
superdanmriedem: so whichever one wins the race I guess16:33
*** sahid has quit IRC16:33
mriedemthere is an issue with yours16:34
mriedemmaybe we should slide https://review.openstack.org/#/c/508299/ in underneath?16:34
superdanah, I didn't see that it called that16:36
*** Yikun__ has quit IRC16:36
superdanlet's just try to get yours in so I don't have to intertwine those two series16:36
superdanstill no jenkins run on that one.. is it worth rechecking now or no?16:36
mriedemno idea16:37
mriedemi don't see it queued up16:37
mriedemi'm going to lunch16:37
*** edmondsw has quit IRC16:37
*** cdent has joined #openstack-nova16:39
openstackgerritEric Fried proposed openstack/nova master: WIP: Use ksa adapter for cinder client  https://review.openstack.org/50834516:43
*** itlinux has quit IRC16:44
*** gyee has joined #openstack-nova16:45
*** Tom has joined #openstack-nova16:45
*** lnxnut_ has joined #openstack-nova16:46
*** itlinux has joined #openstack-nova16:47
*** rtjure has quit IRC16:49
*** Tom has quit IRC16:50
openstackgerritEric Fried proposed openstack/nova master: WIP: Use ksa adapter for cinder client  https://review.openstack.org/50834516:50
cdentmriedem: you aware of this interoperability tag thingie for nova? it wants your +1: https://review.openstack.org/#/c/506255/16:54
*** shaner has quit IRC16:54
*** Swami has joined #openstack-nova17:02
*** itlinux has quit IRC17:02
*** ralonsoh_ has quit IRC17:02
*** mingyu_ has joined #openstack-nova17:03
*** mingyu_ has quit IRC17:03
*** mingyu_ has joined #openstack-nova17:04
*** mingyu has quit IRC17:04
*** gouthamr has quit IRC17:10
*** avolkov has quit IRC17:11
cdentleakypipes: if you’ve got a chance this (shared rp fix from alex) probably wants your eyes: https://review.openstack.org/#/c/480379/17:15
*** jpena is now known as jpena|off17:16
*** yamamoto has joined #openstack-nova17:20
*** mingyu_ has quit IRC17:21
*** mingyu has joined #openstack-nova17:21
*** sri_ has quit IRC17:21
leakypipescdent: ah, yes, thanks for the reminder17:23
cdent17:23
*** yamamoto has quit IRC17:25
*** burt has joined #openstack-nova17:30
*** huanxie has joined #openstack-nova17:31
*** gouthamr has joined #openstack-nova17:33
*** tssurya has quit IRC17:36
*** Tom_ has joined #openstack-nova17:36
*** edmondsw has joined #openstack-nova17:37
*** esberglu has quit IRC17:40
*** mingyu_ has joined #openstack-nova17:40
*** esberglu has joined #openstack-nova17:40
*** esberglu has quit IRC17:41
*** esberglu has joined #openstack-nova17:41
*** dave-mccowan has quit IRC17:41
*** jmlowe has joined #openstack-nova17:41
*** esberglu has quit IRC17:42
*** esberglu has joined #openstack-nova17:42
*** esberglu has quit IRC17:42
*** esberglu has joined #openstack-nova17:42
*** pino has joined #openstack-nova17:43
*** mingyu has quit IRC17:44
*** thorst has quit IRC17:45
pinopenick: mriedem: (and Jay Pipes, can't find his handle) a few days ago I asked about injecting dynamically generated SSH certificates into the host. You pointed me to vendordata, cloud-init (and the novajoin example). I've played with it and have some more questions/issues. Do you have a few minutes?17:46
leakypipespino: I'm leakypipes on Fridays :)17:47
pinoSpecifically, I'm worried about the SSH host certificate.17:47
*** esberglu has quit IRC17:47
pinoleakypipes: :-)17:47
pinoSo, the problem with the host certificate is that you need to pass the host public key to the Host CA signing service to generate the certificate, then inject it back.17:48
pinoBut vendor data doesn't take any arguments, so you're forced to either generate the host keys off the instance, or you need to expose a service directly to the instance.17:49
pinoBoth are problematic: exposing the signing service to the instance means the instance has to be able to authenticate itself to the service (chicken-and-egg problem).17:50
mriedeminterop tag?!17:50
pinoIs that comment for me?17:50
mriedemno17:51
mriedemsorry17:51
*** cdent has quit IRC17:51
pinoMaybe an e-mail would be more effective...17:53
*** rtjure has joined #openstack-nova17:53
pinoI'll write it all up and send to the mailing list instead.17:53
mriedempino: to the openstack-dev mailing yeah, probably17:53
*** lnxnut_ has left #openstack-nova17:55
*** david-lyle has quit IRC17:56
*** david-lyle has joined #openstack-nova17:56
*** gjayavelu has joined #openstack-nova17:57
*** kenperkins has quit IRC17:58
*** mnestratov has quit IRC18:00
*** huanxie has quit IRC18:01
mriedemleakypipes: the explanations for https://review.openstack.org/#/c/506093/ are in the commit message18:02
mriedemthe functional test fails without those changes to the fake virt driver18:02
leakypipesmriedem: I still couldn't understand the explanation.18:02
leakypipesmriedem: which is what I said in the comment../.18:02
mriedemif you do live migration with block_migration=auto,18:02
mriedemthat passes block_migration=None,18:02
mriedemwhich you can't set on MigrateData b/c it's not a nullable=True field18:02
cburgessDo we have a supported/approved method for purge deleted instances from the DB?18:03
mriedemand we have to return migrate_data from the other method otherwise it doesn't come out during post_live_migration, and we can't set the status on the migration object to 'completed', which is what the functional test is waiting for18:03
mriedemcburgess: we don't18:03
cburgessmriedem Wow still really?18:03
mriedemwe had a spec a few releases ago18:03
mriedemand a patch at one point18:03
cburgess*sad panda*18:04
cburgessMy life just got a lot more complicated.18:05
mriedemdid you find out you're pregnant?18:05
cburgessNo the lack of purge support...18:05
mriedemoh, that18:05
melwittI thought archived_deleted_rows mostly worked now18:05
mriedemhttps://specs.openstack.org/openstack/nova-specs/specs/mitaka/approved/purge-deleted-instances-cmd.html18:05
mriedemarchive != purge18:06
cburgessWait..18:06
cburgessOK..18:06
superdanarchive should be a lot better now yeah18:06
cburgessSo whats archived_deleted_rows18:06
mriedemarchive means, put some shit in the storage room but don't delete it18:06
cburgess?18:06
mriedemmoves things to shadow tables18:06
*** hoonetorg has quit IRC18:06
superdanand then you can dump the storage room after each thing18:06
mriedemdeleted things18:06
superdaneasily18:06
melwittit deletes it from the main tables18:06
superdanwithout constraints18:06
cburgessOh..18:06
cburgessThat would do what I needed.18:06
mriedemcburgess: nova-manage db archive_deleted_rows --verbose --until-complete18:06
cburgessCool. I'll look into it. Thanks.18:07
cburgessOh this.. https://blueprints.launchpad.net/nova/+spec/archive-deleted-rows-all18:07
superdanmriedem: wanna peek behind my curtain?18:07
mriedemit's not multicell aware, but you're on folsom so that doesn't matter for you18:07
cburgessLooks like melwitt Worked on it.18:07
cburgessI'm not on folsom.. god..18:07
cburgessWe never used Folsom.18:07
mriedemsuperdan: which curtain?18:07
superdanmriedem: this one over here.. the one that hides my current git tree18:07
melwittyeah, there might be some issues if you keep the archived data indefinitely. recently I learned primary keys can be reused after a DB restart with innodb, so archive can fail if there are dupes18:08
superdanmelwitt: cburgess wants to dump the shadow tables immediately though, so no problem18:08
melwittyep +118:08
cburgessYeah cool.18:08
cburgessI will investigate this feature. Thanks team!18:08
melwittjust sharing my recent revelation, just in case18:08
superdanmriedem:18:09
superdan% gsl HEAD^..18:09
superdana3d7d30 Merge build requests into the sortmaster 900018:09
superdanmriedem: you may ooh and ahh now18:09
mriedemi don't know what you're talking about18:10
mriedemoh18:10
superdanbah18:10
mriedemnow i get it18:10
mriedemjesus, is there anything the sortmaster 9000 can't do?18:11
mriedemoh i know18:11
mriedemRFE: please make the sortmaster 9000 allow me to pass a volume type during boot from volume18:11
mriedemoh btw18:12
mriedemif you're looking for queued stuff18:13
mriedemhttp://zuulv3.openstack.org/18:13
*** rmart04 has joined #openstack-nova18:13
*** rmart04 has quit IRC18:13
mriedemthis pike regression fix needs another +2 https://review.openstack.org/#/c/507938/18:14
mriedemthe zuul failure is the nnet job18:14
superdanwhich hasn't merged yet right?18:15
superdanthe nnet fix?18:15
mriedemi'm watching it18:15
mriedembut no18:16
*** boris_42_ has joined #openstack-nova18:16
melwittI have a tox.ini patch up for adding stderr/stdout capture and test timeout to the functional and api_sample_tests https://review.openstack.org/#/c/50797618:16
melwittif anyone wants to help get stdout capturing back18:17
mriedemmelwitt: commented,18:19
mriedemthere might be a way to avoid the duplication18:19
melwittthanks18:20
*** yamamoto has joined #openstack-nova18:22
*** lpetrut has joined #openstack-nova18:26
*** yamamoto has quit IRC18:26
mriedemok i think these nnet zuul job changes are all passing18:36
mriedemso i'll bug infra after that18:36
*** tonygunk has quit IRC18:38
*** nikhil has joined #openstack-nova18:39
cburgessmriedem melwitt Is there a minimum reliable version that supported the archive thing (as in not before liberty etc)? Looks like the command first went in for grizzly.18:44
cburgesssuperdan too18:44
mriedemlooking18:44
mriedemyou want the --until-complete thing18:44
cburgessOK.18:44
cburgessI can figure out when that went in then.18:44
mriedemi want to say ocata18:44
cburgessPrior to that it didn't work at all or... it was just inefficient?18:45
mriedemhttps://review.openstack.org/#/c/378718/18:45
mriedemyeah ocata18:45
superdandidn't work at all really18:45
mriedembefore that i think there were some random foreign key constraints that could mess it up for totally completing18:45
*** thorst has joined #openstack-nova18:45
mriedemplus you had to always specify the batch number18:45
cburgessRAD...18:46
cburgessSo.. not as sad a panda as I was before, still work for me to do. Thanks.18:46
*** cdent has joined #openstack-nova18:47
mriedemalso https://review.openstack.org/#/q/I77255c77780f0c2b99d59a9c20adecc85335bb1818:47
mriedemso going back to mitaka i think you can delete everything in batches18:47
mriedemper dan's fix18:47
cburgessCool.18:48
cburgessI doubt I'm going to enjoy backporting this to Icehouse. *sigh*18:49
*** cdent has quit IRC18:49
melwitticehouse? ouch18:49
cburgessmelwitt Don't judge.... ok well judge.18:49
melwittnot judging, just imagining the pain18:50
*** thorst has quit IRC18:50
mriedemcburgess: i expect it's actually not bad18:50
mriedemsince that code is isolated18:50
mriedemand not a lot of changes over time18:50
melwittthat reminds me, I wanted to resurrect my old patch to turn on FK constraint enforcement to sqlite18:50
cburgessmelwitt My world is pain.. nothing but pain. I'm a purveyor of fine vintage clouds (actually I purvey new ones, but I have to support vintage ones).18:50
melwitthah18:50
openstackgerritmelanie witt proposed openstack/nova master: Make setenv consistent for functional and api_samples  https://review.openstack.org/50797618:52
openstackgerritmelanie witt proposed openstack/nova master: Make setenv consistent for functional and api-samples  https://review.openstack.org/50797618:54
openstackgerritMatt Riedemann proposed openstack/nova stable/pike: Add a regression test for bug 1718455  https://review.openstack.org/50859018:57
openstackbug 1718455 in OpenStack Compute (nova) "[pike] Nova host disable and Live Migrate all instances fail." [Medium,In progress] https://launchpad.net/bugs/1718455 - Assigned to Matt Riedemann (mriedem)18:57
openstackgerritMatt Riedemann proposed openstack/nova stable/pike: Ensure instance can migrate when launched concurrently  https://review.openstack.org/50859118:57
*** pino has quit IRC19:03
*** lbragstad has quit IRC19:03
*** smatzek has quit IRC19:04
*** smatzek has joined #openstack-nova19:06
*** smatzek has quit IRC19:06
*** lbragstad has joined #openstack-nova19:07
*** pino has joined #openstack-nova19:07
*** penick has quit IRC19:08
mriedemsuperdan: dead code now right? https://github.com/openstack/nova/blob/ae4b5d0147cb3e345bf57034221e9c8fedf3cad2/nova/conductor/manager.py#L26119:09
mriedemoh nvm19:09
mriedemwe're at 1.1719:09
superdanright19:09
superdanwe never bumped19:09
mriedemwe should remember to do that this release, across the board19:10
mriedemwe're due19:10
mriedemand by we i mean you, because i'd just mess it up19:10
superdanheh19:11
*** esberglu has joined #openstack-nova19:12
*** esberglu has quit IRC19:12
*** esberglu has joined #openstack-nova19:13
*** smatzek has joined #openstack-nova19:14
openstackgerritDan Smith proposed openstack/nova master: Merge build requests into the sortmaster 9000  https://review.openstack.org/50859519:16
superdanmriedem: ^ probably needs a few more test cases, but you love pointing those out so I'll indulge you19:16
superdanI'm trying to decide if I should parallelize fetching the buildrequests as well, to try to further lower the latency19:17
superdanit'd be hard to prove that is worth it though, since it's hard to have many instances in that state for long without bugs19:17
*** esberglu has quit IRC19:17
mriedemi'm currently doing a review of https://review.openstack.org/#/c/498948/ before it merges19:18
superdananother favorite pasttime of yours19:18
mriedemyeah the only way to do that really is via fixtures, to hold up the build requests19:18
superdanwell, I meant in a devstack type environment, but yeah19:18
*** sridharg has quit IRC19:19
mriedemyeah i'm not sure how you'd reliably test that19:19
superdanthat's my point yeah19:19
mriedemunless you booted so many at once, and scheduling was slow enough, that you could see a difference19:19
superdanand probably not much gain for people19:19
mriedemsingle node devstack is probably not going to work there19:19
mriedemyou'd have to fake out like 1000 computes or something19:19
mriedemsomething that placement has to chew on19:19
*** catintheroof has quit IRC19:20
superdannah, just shut down scheduler so we block on making that decision and then try to list before the timeout19:20
superdanbut still, not great19:20
*** catintheroof has joined #openstack-nova19:20
*** ijw has quit IRC19:21
mriedemtrue, set the rpc timeout to 10 minutes :)19:21
superdanthe problem is, because we don't have real threads, we're already pegging the cpu just processing the results from the DBs as it is, so, not likely the even smaller overlap of operations is really going to help19:22
melwitthow can we get real threads19:22
superdanI'm not sure eventlet is threadsafe, so I'm not sure we can without bigger changes19:23
*** yamamoto has joined #openstack-nova19:23
superdanthat was the problem with us using real threads for db workers ages ago, and it was never fixed, AFAIK, we just worked around it by moving to an all-python db driver19:23
superdanin aggregate it's not as big of a deal because multiple requests will be running at once in our workers, so more overall work gets done with more cores, just not in a single-request sort of environment19:24
*** catintheroof has quit IRC19:24
superdanany environment that only has one concurrent request ever is either (a) dan's test box or (b) probably not worried about listing thousands of instances at a time :)19:25
melwittheh19:25
*** yamamoto has quit IRC19:28
mriedemsuperdan: some comments/questions in https://review.openstack.org/#/c/498948/19:29
openstackgerritpriyaduggirala proposed openstack/nova master: Rename parameters in call() of nova/image/glance.py  https://review.openstack.org/50853319:29
mriedemnot trying to block, just want to make sure i know what's going on with this stuff before it merges and i'm lost later19:29
superdanmriedem: I thought we didn't do _LI( but we don't do _( at all anymore?19:31
mriedemwe don't translate log messages at all anymore19:32
superdanreally thought I was getting pep8 fails19:32
superdanchrist, I can never keep it straight19:32
mriedemhttps://docs.openstack.org/oslo.i18n/latest/user/guidelines.html19:33
mriedemi think that top paragraph is the new guideline19:34
*** crushil has quit IRC19:34
mriedemand https://docs.openstack.org/oslo.i18n/latest/user/guidelines.html#log-translation19:34
*** thorst has joined #openstack-nova19:35
superdanI believe you, I just can't keep track of it19:37
mriedemi had to look it up too, wasn't sure about _() when you asked but was pretty sure19:37
superdanI figure if I just pick some behavior I'll be right 20% of the time when we've circled back to that as the preferred one19:38
mriedemdepends on the current ibm corporate wide software guidelines at the time19:39
superdanmriedem: so, this isn't in the gate yet so do you want me to fix the bottom patch or tack on to the end? this set is pretty fragile so if I do the bottom it'll likely percolate awesomeness up the stack pretty good19:39
superdanand by "awesomeness" I mean "my tears"19:39
mriedemwas there anything major? the translation markers and simple logging stuff, plus docstring or whatever can all be done at the end19:40
mriedemthere was the one conditional block that i thought was dead code19:40
superdanthe else?19:40
superdanthe comment was just incorrect19:40
mriedemoh19:40
mriedemthe bottom 2 are approved so i'd say just take a fixup change at the end of the series19:41
superdanI really should just remove "on the source" from that one since it'll be used any time we need to revert the allocation regardless of why/where19:41
mriedemthe thing i wanted to be cautious of was overusing it like remove_provider_from_instance_allocation in the scheduler report client,19:42
mriedembecause remove_provider_from_instance_allocation was written really for cold migrate / resize + resize to same host,19:42
mriedembut we've used it for other move operations, and there are some assumptions in there which don't always work for other move operatoins19:42
mriedemi.e.19:42
mriedem# NOTE(danms): We are in a resize to same host scenario. Since we19:42
mriedem            # are the only provider then we need to merge back in the doubled19:42
mriedem            # allocation with our part subtracted19:42
superdanwell, this really should apply to all of the types once we get it done because same/different, cold/live, they're all the same process since we don't have to worry about clashes since the migration uuid holds things19:43
mriedemif so,19:43
mriedemthen heed my warning about assuming migration.source_node is set19:43
mriedemHEED IT19:43
mriedembecause we don't even attempt to set that shit for live migration19:43
superdanyeah, I saw, I need to go look19:43
mriedemsince there is no claim19:43
superdanbecause I thought it was set there too19:43
*** penick has joined #openstack-nova19:44
mriedem_live_migrate in conductor task manager19:44
mriedemwe set the compute attributes, but not the node ones19:44
mriedemwas just looking at using those yesterday for something else19:44
mriedemif those do get set, it would happen later in the compute probably19:44
superdanoh the hosts I see19:44
superdanwell, I expect that just means we don't have coverage for live migrations failing in a way that will result in this getting called,19:46
superdanand/or I wasn't consistent in that patch at the top,19:46
superdanbut I haven't really gone over that one in as much detail19:46
penickpino: There's a couple ways to do this, one is to completely generate a ssh host key and sign it, then stuff it into the instance through vendordata. Another is.. more complex.19:46
superdankeeping these two sets straight has been hard19:47
mriedemsuperdan: i bet19:47
mriedemdon't forget about evacuate and unshelve19:47
mriedemoh yeah, plus the migration object has dest_compute, dest_host, and dest_node, where one of those is the hostname and one is the host IP19:47
mriedemyay!19:47
penickpino: the more complex version means running a whole AuthNG infrastructure to handle trusted boot of an instance and giving it one-time credentials which allow it to bootstrap on boot and get its key signed. Heavy to implement, useful long term for secure infrastructure. Probably way heavier than you need19:48
superdanmriedem: not sure unshelve is a thing here.. once you're offloaded we shouldn't keep an allocation for you19:48
mriedemoh right19:49
mriedemyeah19:49
superdanand for evacuate, we don't need to keep a claim on the old host, so ... also not sure that's a thing19:49
superdanI originally had them on my radar, but thinking more I'm not sure that's the right thing to do19:49
* superdan sighs looking at the gate19:52
*** mingyu_ has quit IRC19:58
*** mingyu has joined #openstack-nova19:58
*** smatzek has quit IRC19:59
*** smatzek has joined #openstack-nova20:00
*** huanxie has joined #openstack-nova20:00
*** MikeG451 has quit IRC20:01
dimsmriedem : if i can get to the browser based vnc for a nova vm ... is there a way to attach a real vncviewer instead of the browser? (is there enough information in the browser url?)20:02
*** chyka_ has joined #openstack-nova20:02
superdandims: you need a websocket client proxy20:03
superdandims: the browser client isn't connecting to the vnc server, but getting the stream over a websocket20:03
dimssuperdan : i see, i have to find a vnc client that supports that ..20:04
dimsthanks superdan. will look20:04
superdandims: it's called novnc :P20:04
dimswhich is browser based :)20:04
*** smatzek_ has joined #openstack-nova20:04
superdanright20:04
superdanI expect you won't find another, but .. good luck :)20:05
*** smatzek has quit IRC20:05
superdanyou could build a ws proxy that listens on a tcp port20:05
superdanclient->proxy->proxy->server20:05
*** chyka has quit IRC20:05
dimshmm, i might just stick to the browser :)20:07
superdangood plan :)20:07
*** esberglu has joined #openstack-nova20:07
*** tonygunk has joined #openstack-nova20:08
*** smatzek_ has quit IRC20:09
*** xyang1 has quit IRC20:11
*** esberglu has quit IRC20:11
mriedemdims: are you trying to circumvent zones20:12
dimsmriedem : not really, i can use the browser on my mac already, its just not very convenient for long use20:13
mriedems/*/i can use my mac already, it's just not very convenient/20:14
dimsmriedem : bad regex? :)20:15
mriedemyeah maybe20:15
superdandims: vnc consoles are for disaster recovery and windows boxes, neither of which should be prolonged use20:16
dimsright superdan. there's just no other way to get to these vm(s) currently (multi-level protected zones)20:17
superdanouch20:17
dimsy sigh20:17
*** syjulian has quit IRC20:17
openstackgerritMatt Riedemann proposed openstack/nova master: Add hints to what the Migration attribute values are  https://review.openstack.org/50860320:21
openstackgerritMatt Riedemann proposed openstack/nova master: Refactor duplicate code for looking up the compute node name  https://review.openstack.org/50860420:21
*** mnestratov has joined #openstack-nova20:22
*** yamamoto has joined #openstack-nova20:24
cfriesennot strictly a nova question, but hoping someone knows.  If I monkey_patch the world with eventlet, will threading.Thread() actually give me a real thread or will it give me a greenthread?20:25
superdana greenthread20:27
*** tonygunk has quit IRC20:27
*** yamamoto has quit IRC20:29
cfriesenthanks.  I was expecting a real thread and couldn't figure out why I didn't see it on the system.20:29
*** huanxie has quit IRC20:30
cfriesendo Condition and RLock and friends all work properly with the greenthread?20:32
superdanyes20:32
*** syjulian has joined #openstack-nova20:36
*** gjayavelu has quit IRC20:36
*** MikeG451 has joined #openstack-nova20:36
openstackgerritEd Leafe proposed openstack/nova-specs master: Return Alternate Hosts  https://review.openstack.org/50427520:43
*** slaweq has joined #openstack-nova20:47
*** nikhil has quit IRC20:49
*** Sukhdev has joined #openstack-nova20:49
*** jmlowe has quit IRC20:50
*** bnemec has quit IRC20:51
*** vladikr has quit IRC20:55
*** armax has joined #openstack-nova21:00
mriedemwhat was the big reason why baremetal needs to have reschedules again? came up in boston and the ironic multinode jobs rely on reschedules right now - is it something with node contention? or was it also something with just failed hardware builds?21:04
superdanflakiness of the management controllers21:05
*** mnestratov has quit IRC21:05
superdanbefore claims in the scheduler, it was also that only one instance can ever be there,21:05
superdanso if we race for resources with virt, there is a possibility that both instances can fit, but that never happens with bm21:06
*** awaugama has quit IRC21:06
*** mingyu has quit IRC21:09
*** penick has quit IRC21:10
*** thorst has quit IRC21:10
*** wxy has quit IRC21:13
*** jamesdenton has joined #openstack-nova21:17
*** dgonzalez_ has joined #openstack-nova21:19
*** sapcc-bot has quit IRC21:19
*** gjayavelu has joined #openstack-nova21:19
*** sapcc-bot1 has joined #openstack-nova21:19
*** edmondsw has quit IRC21:20
*** jpena|off has quit IRC21:21
*** dgonzalez_ has quit IRC21:21
mriedemfigleaf: some comments inline https://review.openstack.org/#/c/504275/721:24
mriedembut overall good21:24
mriedemsuperdan: ok, that was something i mentioned in ^ as a missing use case21:24
mriedemsince it seemed like a pretty critical issue for ironic21:24
*** ltomasbo has quit IRC21:24
*** yamamoto has joined #openstack-nova21:25
*** slaweq has quit IRC21:26
*** slaweq has joined #openstack-nova21:26
*** ltomasbo has joined #openstack-nova21:28
*** jpena|off has joined #openstack-nova21:29
openstackgerritayoung proposed openstack/nova master: Admin API Policy contingent on is_admin_project  https://review.openstack.org/38414821:29
*** thorst has joined #openstack-nova21:30
*** slaweq has quit IRC21:31
*** yamamoto has quit IRC21:31
*** slaweq has joined #openstack-nova21:34
*** baoli has quit IRC21:35
*** thorst has quit IRC21:35
*** hemna_ has quit IRC21:38
mriedemleakypipes: fried_rice: is there a reason why the traits parameter in alex's spec for GET /allocation_candidates is just called "required" rather than "required_traits"?21:41
mriedem"Is the 'required' parameter required? No, it's optional. OK?"21:41
figleafmriedem: thanks21:43
figleafmriedem: and I wasn't at Boston, but I remember it from Atlanta21:45
mriedemok21:45
mriedemlet's combine our forgetfulness21:45
mriedemfor max recollection21:45
figleafIt was in Austin!!21:46
*** sapcc-bot1 has quit IRC21:47
*** sapcc-bot has joined #openstack-nova21:47
*** hoonetorg has joined #openstack-nova21:50
*** huanxie has joined #openstack-nova21:50
*** esberglu has joined #openstack-nova21:50
mriedemleakypipes: figleaf: superdan: johnthetubaguy: fried_rice: left some comments in the traits spec https://review.openstack.org/#/c/497713/ - two things i don't like, but i'm assuming have been discussed to death already21:50
superdanI won't get back to it until monday21:51
superdanbut... cool21:51
figleafmriedem: I don't like 'required' or 'requires' either21:51
superdanrequired= is the only way to go, IMHO21:52
*** thorst has joined #openstack-nova21:52
figleafmriedem: heading out soon; will comment by Monday21:52
*** pchavva has quit IRC21:54
*** esberglu has quit IRC21:56
*** figleaf is now known as edleafe21:56
*** thorst has quit IRC21:56
*** lpetrut has quit IRC21:58
*** shaner has joined #openstack-nova21:58
mriedem?required_foobars21:58
*** ijw has joined #openstack-nova22:01
*** Swami has quit IRC22:01
*** ijw has quit IRC22:02
mriedemok seeing the bike shedding on this in the previous comments22:05
*** lyan has quit IRC22:06
fried_ricemriedem I abstain.  I'd be fine with whatever y'all land on wrt those two points.22:06
*** slaweq has quit IRC22:08
*** slaweq has joined #openstack-nova22:15
*** huanxie has quit IRC22:20
*** armax has quit IRC22:26
*** yamamoto has joined #openstack-nova22:27
mriedemtested out GET /servers?status=''22:28
mriedemit definitely has a behavior change from what the spec is proposing22:28
mriedemit applies that as the filter, which means you don't get any instances back that have a vm_state set22:28
*** gouthamr has quit IRC22:29
*** ijw has joined #openstack-nova22:31
*** baoli has joined #openstack-nova22:32
*** yamamoto has quit IRC22:32
*** baoli_ has joined #openstack-nova22:34
*** baoli has quit IRC22:37
fried_ricemriedem But does status= take a list?22:41
fried_riceIt doesn't appear to, in my trials.22:41
fried_ricemriedem Which makes it apples:oranges compared to traits.22:41
thingeemriedem: there are 69 total forum sessions22:44
thingeemriedem: was waiting all day for that answer myself :)22:44
*** esberglu has joined #openstack-nova22:45
mriedemok and 73 proposed22:45
mriedemso i guess chances are good to get stuff in, especially since there are some duplicates22:46
thingeehey I'll cheers my beer to that22:46
*** slaweq has quit IRC22:48
*** esberglu has quit IRC22:50
*** thorst has joined #openstack-nova22:53
*** pino has quit IRC23:04
*** slaweq has joined #openstack-nova23:04
*** chyka_ has quit IRC23:05
*** hongbin has quit IRC23:09
*** huanxie has joined #openstack-nova23:20
openstackgerritEric Fried proposed openstack/nova master: Use ksa adapter for cinder client  https://review.openstack.org/50834523:21
fried_ricemriedem Since you're the cinder guy? ^23:22
*** sapcc-bot has quit IRC23:23
*** sapcc-bot has joined #openstack-nova23:23
*** yamamoto has joined #openstack-nova23:29
*** yamamoto has quit IRC23:35
*** lbragstad has quit IRC23:35
*** slaweq has quit IRC23:36
superdanmriedem: they just restarted zuul with some fixes including one that makes it not do the thing your patch was trying to avoid23:37
superdanso we might be unblocked23:37
*** esberglu has joined #openstack-nova23:39
*** fried_rice is now known as efried_thbagh23:40
*** esberglu has quit IRC23:44
*** slaweq has joined #openstack-nova23:50
superdanmriedem: strike that.23:51
*** zhurong has joined #openstack-nova23:52
*** huanxie has quit IRC23:53
*** gyee has quit IRC23:53
*** Sukhdev has quit IRC23:59

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!