Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-02-16
10:08:41 hrw this one adds test and new config option
10:11:01 stelucz_ Hello, is there nova client command to delete resource_provider from database? After running nova service-delete <id>, record for compute node still exists in resource_providers table, thus reprovisioning of compute node ends up in message: Another thread already created a resource provider with the UUID 7b01cc27-c101-4c05-aaed-5958ef1270a1. Grabbing that record from the placement API.
11:15:17 openstackgerrit Merged openstack/nova master: api-ref: provide more detail on what a provider aggregate is https://review.openstack.org/539033
11:45:08 openstackgerrit Marcin Juszkiewicz proposed openstack/nova master: Allow to configure amount of PCIe ports in aarch64 instance https://review.openstack.org/545034
11:48:41 openstackgerrit Stephen Finucane proposed openstack/nova master: trivial: Move __init__ function https://review.openstack.org/538223
11:48:41 openstackgerrit Stephen Finucane proposed openstack/nova master: tox: Add mypy target https://review.openstack.org/538221
11:48:42 openstackgerrit Stephen Finucane proposed openstack/nova master: tox: Store list of converted files https://review.openstack.org/538222
11:48:42 openstackgerrit Stephen Finucane proposed openstack/nova master: mypy: Add type annotations to 'nova.pci' https://review.openstack.org/538224
11:48:43 openstackgerrit Stephen Finucane proposed openstack/nova master: zuul: Add 'mypy' job https://review.openstack.org/539168
12:26:40 larsks Hey nova folk: there are some libvirt guests on our pike compute nodes that have an empty <nova:owner> attribute in the libvirt xml (which is causing ceilometer to blow up). Is an empty owner a known problem? Or is that expected behavior in some circumstances?
12:27:05 hrw larsks: maybe owner account got removed from system?
12:28:30 larsks hrw: the owner is the associated project, right? The project still exists.
12:28:54 larsks And would nova update the xml on a running guest like that in any case?
12:29:17 hrw <nova:owner><nova:user /><nova:project /></nova:owner>
12:29:51 hrw larsks: sorry, no idea. I jsut hack nova
12:30:08 hrw and use bigger and bigger knives each time
12:30:33 larsks Fair enough :).
12:49:34 cdent finucannot: has mypy found some juicy bugs in nova yet?
13:32:42 hrw anyone know how to get is guest instance i440fx or q35?
13:35:59 hrw found. guest.os_mach_type
13:37:28 finucannot cdent: Nothing exciting yet but it will as it expands. I started with 'nova.pci' simply because comprehending that stuff is currently a nightmare and types help
13:38:04 cdent Will be cool to see how things play out
14:01:08 openstackgerrit Marcin Juszkiewicz proposed openstack/nova master: Allow to configure amount of PCIe ports in aarch64 instance https://review.openstack.org/545034
14:01:21 hrw one step closer
14:01:29 hrw shit. have to fix commit message now
14:02:11 openstackgerrit Marcin Juszkiewicz proposed openstack/nova master: Allow to configure amount of PCIe ports https://review.openstack.org/545034
14:06:37 openstackgerrit Eric Young proposed openstack/nova master: Enable native mode for ScaleIO volumes https://review.openstack.org/545304
14:07:24 hrw one simple thing, few projects involved to find out what is going on. nova, libvirt, qemu, uefi :(
14:30:26 mnaser this simple clean up has been a lot messier than expected :(
14:31:20 mnaser melwitt: for some reason that example you have listed has multiple attachments so that’s probably why it still is marked as reserved? mriedem maybe can confirm this?
14:36:25 finucannot bauwser: Could you knock this one through? https://review.openstack.org/#/c/538223/
14:37:08 fried_rice stelucz: I believe osc has the command you're looking for. But it doesn't sound like you need to delete it in that scenario. The message you're seeing is not an error.
14:54:22 melwitt mnaser: multiple attachments? no
14:55:06 lyarwood sean-k-mooney: https://bugs.launchpad.net/os-vif/+bug/1749972 - finucannot pointed me in your direction regarding this odd `brctl setageing $bridge 0` behaviour I've just stumbled across with the latest 16.04 kernel, would you mind taking a look?
14:55:07 openstack Launchpad bug 1749972 in os-vif "`brctl setageing $bridge 0` fails on Ubuntu 16.04 4.4.0-21-generic" [Undecided,New]
14:55:17 mnaser melwitt: in your curl example, I saw multiple attachments returned
14:55:30 mnaser I count 4
14:55:47 mnaser (To 4 différence instances)
14:55:55 mnaser Different*
14:56:16 openstackgerrit Eric Fried proposed openstack/nova master: api-ref: Further clarify placement aggregates https://review.openstack.org/545356
14:56:23 fried_rice figleaf: ^^^
14:56:52 melwitt mnaser: oh, weird. I didn't notice that
14:56:55 fried_rice mriedem, superdan: bauwser also ^
14:57:44 mnaser melwitt: I can only guess that maybe you tried to run this before applying patch so a few got accumulated? Or maybe it’s a multiattach volume?
14:57:49 melwitt I had been reusing the same volume as I messed up a bunch of tests. I'll try again with a new one
14:58:01 bauwser fried_rice: +2
14:58:11 fried_rice thx
14:58:21 bauwser we can iterate long about docs :)
14:58:33 bauwser but I'm fine
15:00:39 fried_rice As long as each iteration is an improvement over the previous, agree that we need not make every patch perfect.
15:04:01 melwitt mnaser: indeed you're right, it works fine with a fresh volume with only one attachment. thanks for pointing that out
15:04:20 melwitt I'll respin it to remove that code comment I wrote that was wrong
15:04:43 mnaser melwitt: awesome!
15:16:06 melwitt correction, with a fresh volume it actually has no attachments, if I make it fail to build on compute. it ends up 'reserved' with no attachments, then the detach makes it 'available' then the delete deletes the volume
15:17:56 openstackgerrit melanie witt proposed openstack/nova master: Clean up ports and volumes when deleting ERROR instance https://review.openstack.org/340614
15:17:57 openstackgerrit melanie witt proposed openstack/nova master: Add functional recreate test of deleting a BFV server pre-scheduling https://review.openstack.org/545123
15:17:57 openstackgerrit melanie witt proposed openstack/nova master: Detach volumes when deleting a BFV server pre-scheduling https://review.openstack.org/545132
15:18:52 tobasco i have a ongoing issue right now, nova metadata api returns error code 300 to neutron-metadata-agent, so cloud-init fails, neutron metadata givens error 500 to cloud-init
15:19:18 tobasco mitaka, anybody know what would cause a 300 error code? i'm gonna dump the request and try manually now
15:31:32 mriedem it sucks that we have to provide a token to list the versions for the compute api
15:39:31 fried_rice mriedem: We fixed that for placement.
15:39:41 fried_rice We should fix it for nova too
15:40:13 fried_rice mriedem: https://review.openstack.org/#/c/522002/
15:40:37 fried_rice Because the root URI is supposed to be auth-less. mordred will tell you so.
15:40:48 mriedem yeah i remember
15:41:00 superdan mordred will tell you lots of things
15:41:10 fried_rice It just makes sense too though.
15:41:25 fried_rice superdan: Are you saying mordred tends to the garrulous?
15:42:00 superdan oh I dunno about that
15:42:15 superdan he's just mordredulous
15:42:26 fried_rice mordredoquacious
15:44:27 mriedem huh, this is kind of fun
15:44:28 mriedem stack@queens:~$ openstack server list
15:44:28 mriedem +--------------------------------------+---------+--------+----------+-------+--------+
15:44:29 mriedem | ID | Name | Status | Networks | Image | Flavor |
15:44:29 mriedem | fd20384d-c0e5-40ab-b9f5-c3fae406379b | server1 | ERROR | | | |
15:44:29 mriedem +--------------------------------------+---------+--------+----------+-------+--------+
15:44:35 mriedem this server failed in the api i think
15:44:38 mriedem on create
15:44:53 mriedem it's volume-backed so that's why no image, but no flavor?
15:45:50 mriedem melwitt: mnaser: heh, as a user, i just hit the bug we're trying to fix,
15:46:09 mriedem volume-backed server create failed, i delete it, then went to do it again but the volume is reserved
15:46:15 mriedem so now i have to switch to admin to force detach it
15:46:40 melwitt that bug so hot right now
15:49:28 openstackgerrit Marcin Juszkiewicz proposed openstack/nova master: Allow to configure amount of PCIe ports https://review.openstack.org/545034
15:52:01 cdent I think a t-shirt with "that bug so hot right now" would work pretty well. It rings.
15:54:41 mordred superdan, mriedem, fried_rice: yes please to not having auth on discovery urls
15:55:33 fried_rice cdent: Picture of a cockroach with a smug grin and a pompadour
15:56:02 ingy mordred: hey o/
15:56:16 cdent fried_rice: that will do nicely
16:00:01 finucannot FYI, I'm gone for the next week so if anyone pings me don't expect me to answer. See everyone at the PTG!
16:00:11 ingy cdent is OnIt™
16:00:37 cdent ingy: always
16:01:03 openstackgerrit Stephen Finucane proposed openstack/nova master: Rename '_numa_get_constraints_XXX' functions https://review.openstack.org/385072
16:01:03 openstackgerrit Stephen Finucane proposed openstack/nova master: Standardize '_get_XXX_constraints' functions https://review.openstack.org/385071
16:05:16 mriedem am i making some obvious mistake here?
16:05:17 mriedem curl -d '{"os-force_detach": {}}' -H "accept: application/json" -H "x-auth-token: $token" http://199.204.45.19/volume/v3/e9d773beeef2435eb59f7c6eeaf685a9/volumes/126c8d4b-c582-484a-8c09-fe901a7dc17f/action
16:05:17 mriedem {"badRequest": {"message": "There is no such action: None", "code": 400}}
16:05:21 mriedem https://developer.openstack.org/api-ref/block-storage/v3/#force-detach-a-volume
16:06:38 melwitt did you include a request body?

Earlier   Later