| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-04-20 | |||
| 00:27:10 | kevinbenton | well we also wouldn't see it in single node gate | |
| 00:27:20 | kevinbenton | because dhcp agent would get attached before VM | |
| 00:27:29 | kevinbenton | and DHCP agent plugs using correct MTU | |
| 00:27:48 | clarkb | kevinbenton: right but until we test with jumbo frames we would never see it regardless of multinode or single node | |
| 00:27:51 | clarkb | (and we have multinode testing) | |
| 00:27:58 | imacdonn | FWIW, the dhcp_agent's tap interface does get set to that of the network .. but that's obviously nothing to do with nova | |
| 00:28:45 | kevinbenton | so yeah, there are two possible fixes i can see | |
| 00:28:53 | kevinbenton | tell libvirt about MTU if it supports that | |
| 00:29:04 | imacdonn | yeah, if | |
| 00:29:10 | kevinbenton | or proceed with https://review.openstack.org/#/c/447150/ | |
| 00:29:25 | kevinbenton | which will stop libvirt from plugging stuff into linux bridges | |
| 00:29:35 | kevinbenton | and then re-add something like https://git.openstack.org/cgit/openstack/neutron/commit/?id=d352661c56d5f03713e615b7e0c2c9c8688e0132 | |
| 00:29:45 | imacdonn | hmm ineresting | |
| 00:29:46 | kevinbenton | to set MTU before it gets added to the bridge | |
| 00:29:48 | imacdonn | interesting* | |
| 00:29:52 | openstackgerrit | Takashi NATSUME proposed openstack/nova master: Avoid forcing translation on logging calls https://review.openstack.org/413876 | |
| 00:31:55 | imacdonn | interesting hackaround ... http://linuxaleph.blogspot.com/2013/01/how-to-network-jumbo-frames-to-kvm-guest.html | |
| 00:33:27 | kevinbenton | imacdonn: can you check if libvirt can take MTU as an option? | |
| 00:33:35 | imacdonn | Googling on that now | |
| 00:34:02 | kevinbenton | because the problem with my patch is that it requires a permissions change due to qemu sillyness | |
| 00:34:20 | kevinbenton | it requires more permissions to not create a bridge than to create one :) | |
| 00:34:26 | imacdonn | heh | |
| 00:34:49 | clarkb | kevinbenton: well qemu-bridge-helper is a thing, maybe you just make that nop? | |
| 00:34:57 | clarkb | (thought that might confuse people greatly) | |
| 00:35:48 | kevinbenton | clarkb: yeah, that also makes detection harder because i wouldn't be using a different vif_type on the neutron side | |
| 00:36:18 | kevinbenton | it would be hard to distinguish in the agent if a given nova port was going to try to plug itself into a bridge | |
| 00:36:38 | clarkb | gotcha | |
| 00:37:04 | clarkb | (you only need to attach to the bridge if on a network connected to something?) | |
| 00:37:27 | imacdonn | https://libvirt.org/formatnetwork.html <- trying to comprehend this .. I think it's saying that it only works for libvirt-managed networks | |
| 00:38:10 | imacdonn | "Network types not specifically mentioned here don't support having an MTU set in the libvirt network config." | |
| 00:38:58 | kevinbenton | clarkb: we attach ports to bridges that were created by agents. we leave attaching to bridge up to nova for compute ports | |
| 00:39:22 | kevinbenton | clarkb: so we skip based on device_owner 'compute' right now | |
| 00:39:42 | kevinbenton | clarkb: ideally i want to get rid of that, but for one cycle i will need to know about old compute ports trying to plug themselves in | |
| 00:40:32 | kevinbenton | clarkb: so with my current approach we just set the vif_type='tap' for any newly bound linux bridge ports | |
| 00:40:48 | kevinbenton | clarkb: and that will make it easy to tell | |
| 00:41:38 | clarkb | and is there a reason you can't set the mtu on the bridge when it is created (because you don't know what the mtu of the devices that attach to it later will be?) | |
| 00:42:08 | kevinbenton | clarkb: we could | |
| 00:42:11 | imacdonn | linux doesn't seem to allow that .. at least the kernel I'm running | |
| 00:42:13 | kevinbenton | clarkb: but that leads to the other bug | |
| 00:42:28 | kevinbenton | clarkb: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1399064 | |
| 00:42:28 | openstack | Launchpad bug 1399064 in linux (Ubuntu) "bridges cannot have a mtu > 1500 by themselves" [Medium,Triaged] | |
| 00:42:35 | clarkb | ah | |
| 00:42:50 | clarkb | we set them lower in the multinode tests when we used linux bridges and not ovs | |
| 00:42:53 | clarkb | good to know thanks | |
| 00:45:20 | kevinbenton | imacdonn: yeah, that doesn't sound promising for libvirt | |
| 00:45:28 | imacdonn | ack | |
| 00:46:16 | kevinbenton | imacdonn: i suppose we can put the logic back in the neutron agent | |
| 00:46:30 | kevinbenton | the tap device inherits from the bridge, right? | |
| 00:46:44 | imacdonn | it appears so, yes | |
| 00:47:01 | imacdonn | well... | |
| 00:47:08 | kevinbenton | so once the bridge is embiggened, we don't reduce by adding a tap device | |
| 00:47:12 | imacdonn | it picks up the bridge's MTU when it gets added to the bridge | |
| 00:47:12 | kevinbenton | :) | |
| 00:47:21 | imacdonn | if you change the bridge later, I don't think the tap will follow | |
| 00:47:27 | kevinbenton | yeah, that's fine | |
| 00:47:37 | kevinbenton | i'm just trying to make sure the bridge wouldn't shrink every time | |
| 00:47:44 | imacdonn | right | |
| 00:47:47 | kevinbenton | until the agent went through and corrected | |
| 00:47:58 | kevinbenton | ok, then we can fix this in neutron | |
| 00:48:07 | kevinbenton | imacdonn: if you file a bug i can revert that patch | |
| 00:48:18 | kevinbenton | well not revert | |
| 00:48:26 | kevinbenton | but put in logic to match to the network MTU | |
| 00:48:59 | imacdonn | hang on | |
| 00:49:09 | imacdonn | I might have been looking at the wrong libvirt doc | |
| 00:49:26 | imacdonn | https://libvirt.org/formatdomain.html#mtu | |
| 00:49:55 | imacdonn | this is fairly new .. I'm on libvirt 1.2.x | |
| 00:50:28 | imacdonn | well, a mix of 1.2.x and 2.0.x, but not 3.1 | |
| 00:52:37 | kevinbenton | i gtg for a bit. if it's not looking good for libvirt, just let me know | |
| 00:52:51 | kevinbenton | the logic should probably exist in linuxbridge agent anyway to deal with older ones | |
| 00:52:52 | imacdonn | I'll file the bug anyway .. think we at least need something to get us by until your other solution | |
| 00:53:23 | imacdonn | yeah, I think that requiring libvirt 3.1.x would be a problem for many | |
| 00:53:45 | imacdonn | thanks! I'll get the bug in | |
| 01:08:07 | imacdonn | https://bugs.launchpad.net/neutron/+bug/1684326 | |
| 01:08:07 | openstack | Launchpad bug 1684326 in neutron "MTU not set on nova instance's vif_type=bridge tap interface" [Undecided,New] | |
| 01:19:30 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Deprecate os-hosts API https://review.openstack.org/456504 | |
| 01:49:28 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Support tag instances when boot https://review.openstack.org/394321 | |
| 02:11:57 | openstackgerrit | zhao mingjun proposed openstack/nova master: follow-up https://review.openstack.org/458298 | |
| 02:59:19 | openstackgerrit | Huan Xie proposed openstack/nova master: WIP: XenAPI use os-xenapi v2 in nova https://review.openstack.org/453493 | |
| 03:06:50 | openstackgerrit | Huan Xie proposed openstack/nova master: WIP: XenAPI use os-xenapi v2 in nova https://review.openstack.org/453493 | |
| 03:18:52 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Support tag instances when boot https://review.openstack.org/394321 | |
| 04:02:28 | Shashi | Hello Team, | |
| 04:02:33 | openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Support tag instances when boot https://review.openstack.org/394321 | |
| 04:03:31 | Shashi | Getting AttributeError: 'module' object has no attribute 'VIR_MIGRATE_POSTCOPY' error and n-cpu services are not starting up | |
| 04:12:43 | imacdonn | Shashi, if this is a deployment issue, better to ask in #openstack ... but also see https://bugs.launchpad.net/openstack-devel/+bug/1682692 | |
| 04:12:43 | openstack | Launchpad bug 1682692 in OpenStack Development Infrastructure "nova fails with error: 'module' object has no attribute 'VIR_MIGRATE_POSTCOPY' after using libvirt 2.5.0 from UCA" [Undecided,New] | |
| 04:15:24 | Shashi | Thanks imacdonn | |
| 05:07:13 | sj_ | what's the concept of cells in nova ? | |
| 05:07:21 | sj_ | m not getting it | |
| 05:07:26 | sj_ | ny1 ? | |
| 05:08:34 | sj_ | ?? | |
| 05:51:32 | openstackgerrit | Huan Xie proposed openstack/nova master: WIP: XenAPI use os-xenapi v2 in nova https://review.openstack.org/453493 | |
| 06:26:34 | openstackgerrit | Alex Xu proposed openstack/nova master: Deprecate Multinic, floatingip action and os-virtual-interface API https://review.openstack.org/457181 | |
| 06:27:05 | openstackgerrit | Lei Zhang proposed openstack/nova master: Add sync traits command for placement https://review.openstack.org/450125 | |
| 06:28:46 | openstackgerrit | Huan Xie proposed openstack/nova master: WIP: XenAPI use os-xenapi v2 in nova https://review.openstack.org/453493 | |
| 06:33:22 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for flavors-access endpoint instead of stevedore https://review.openstack.org/456589 | |
| 06:33:22 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for '/servers' endpoint instead of stevedore https://review.openstack.org/445864 | |
| 06:33:23 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for flavor endpoint instead of stevedore[1] https://review.openstack.org/456549 | |
| 06:33:23 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for '/os-aggregates' endpoint instead of stevedore https://review.openstack.org/456825 | |
| 06:33:24 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for flavors-extraspecs endpoint instead of stevedore https://review.openstack.org/456585 | |
| 06:33:24 | openstackgerrit | Alex Xu proposed openstack/nova master: Use plain routes list for '/os-keypairs' endpoint instead of stevedore https://review.openstack.org/456478 | |