Wednesday, 2026-07-29

opendevreviewGoutham Pacha Ravi proposed openstack/nova master: Support cold migration with virtiofs shares  https://review.opendev.org/c/openstack/nova/+/98963306:16
opendevreviewGoutham Pacha Ravi proposed openstack/nova master: Allow rebuild for instances with virtiofs shares  https://review.opendev.org/c/openstack/nova/+/99904806:52
opendevreviewGoutham Pacha Ravi proposed openstack/nova master: Allow rebuild for instances with virtiofs shares  https://review.opendev.org/c/openstack/nova/+/99904806:53
*** zseguin is now known as Guest1435007:21
opendevreviewJoan Gilabert proposed openstack/nova master: Report OWNER_NOVA trait on resource providers  https://review.opendev.org/c/openstack/nova/+/99429909:48
opendevreviewJoan Gilabert proposed openstack/nova master: Add libvirt driver MDEV ARQ support for Cyborg  https://review.opendev.org/c/openstack/nova/+/99584009:48
*** ralonsoh_ is now known as ralonsoh11:01
opendevreviewBalazs Gibizer proposed openstack/nova master: Make functional-py313-threading voting  https://review.opendev.org/c/openstack/nova/+/99721613:48
opendevreviewBalazs Gibizer proposed openstack/nova master: Stabilize functional test for slow instance.save  https://review.opendev.org/c/openstack/nova/+/99912113:48
gibiOK now that ^^ series feels complete to stabilize the functional-py313-threading tox target13:49
opendevreviewJoan Gilabert proposed openstack/nova master: Report OWNER_NOVA trait on resource providers  https://review.opendev.org/c/openstack/nova/+/99429913:50
opendevreviewJoan Gilabert proposed openstack/nova master: Add libvirt driver MDEV ARQ support for Cyborg  https://review.opendev.org/c/openstack/nova/+/99584013:50
ralonsohsean-k-mooney, gibi hi folks, I've seen recently a raise in the live-migrations errors, in particular in the ML2/OVS jobs15:05
ralonsohI've reported https://bugs.launchpad.net/nova/+bug/216189415:05
ralonsohcan you please take a look? Because this is happening randomly. Most of the times is because the port ofport=-115:06
sean-k-mooneyinteresitng 15:06
sean-k-mooneyyes we can take a look15:06
ralonsohwith os_vif_delegation=True, that means this is created by qemu15:06
ralonsoh(if I'm not wrong)15:06
sean-k-mooneyyes and no15:06
ralonsohI thought initially that could be related to the tap creation flag, but that was for ML2/OVN only (at least now)15:07
sean-k-mooneyso ml2/ovs is require to finish bidnign prot with of-port=-115:07
sean-k-mooneyos_vif_delegation=True menas we use interface type ethernet15:07
ralonsohyes, but who creates the TAP?15:08
sean-k-mooneylibvirt15:08
ralonsohthat's the point, the TAP is created when the VM is unpaused15:08
sean-k-mooneyif you want os-vif to create the tap then you need to tell os-vif ot do that with the new fucntioantly15:08
ralonsohor at least the ofport assigned15:08
sean-k-mooneyralonsoh: yes15:09
sean-k-mooneyand ml2 ovs is not allowed to rely on the ofport being set15:09
sean-k-mooneyto send network-vif-plugged15:09
ralonsohyes, OVS agent needs to see ofport different from INVALID (-1)15:09
ralonsohand then Neutron API will send the vif-plugged event15:10
sean-k-mooneythat is a neutron bug15:10
sean-k-mooneyti is not allowed to requrie that15:10
sean-k-mooneybecause it will break dpdk if it does15:10
ralonsohthis is why in dpdk we disable live_migration_wait_for_vif_plug15:10
sean-k-mooneyno15:11
ralonsohlet's focus on the kernel datapath15:11
sean-k-mooneyfor dpdk we shoudl not reuire that15:11
sean-k-mooneyok well the ovs agent regardless of the kerle or dpdk15:11
sean-k-mooneyis not allowed to depen on osprot being set to a non -1 value15:12
sean-k-mooneyhttps://github.com/openstack/neutron/commit/7ebcce0b8dcd4cd482da92672f5ae7c51530a42915:12
sean-k-mooneyis likely a regression15:12
sean-k-mooneywell maybe not15:12
sean-k-mooney``network-vif-plugged`` shoudl be sent when osport is -115:13
ralonsohwe have never done this15:13
sean-k-mooneywe never depened on the ofport out side of the openflow security group driver15:14
sean-k-mooneyso if you use ip tbales ofport was never reuqired15:14
sean-k-mooneyand even with the ovs firewall driver we just defered the security gorup flwo setup15:15
ralonsohbut regardless of the FW setup, that is done after the port receives a valid ofport15:15
sean-k-mooneyit wasnt in the past15:16
ralonsohI pushed this https://review.opendev.org/c/openstack/neutron/+/992423 1 month ago15:16
sean-k-mooneythe tap is created befor ethe vm is unpased by the way15:16
ralonsoh(related to the previous patch)15:16
sean-k-mooneybut not a lot before it15:16
ralonsohyes, but the TAP is created by libvirt, and that requires the vif-plugged event15:17
ralonsohI really don't know how this could work now15:18
sean-k-mooneyno the vif plugged event does not depend on the tap exsiting in general15:18
ralonsohwithout the TAP, the ofport=-115:18
ralonsohright?15:18
sean-k-mooneyyes15:18
sean-k-mooneyand that is valid15:18
ralonsohso yes, it depends15:18
ralonsohuntil we have a valid ofport, Neutron does not send vif-plugged15:19
sean-k-mooneyml2/ovs need to supprot seting network-vif-plugged with ofport=-115:19
sean-k-mooneythen that is a regression in neutron15:19
ralonsohwe never did that before15:19
sean-k-mooneyyes we di15:19
ralonsohthis could not be a regression15:19
sean-k-mooneyfor a very very very long time15:19
ralonsohno no, hold on15:20
sean-k-mooneythis is how dpdk has worked for ever15:20
ralonsohI really need to understand how it could work now15:20
ralonsohbecause is working randonly, most of the times15:20
sean-k-mooneyhttps://github.com/openstack/neutron/commit/7fd2725cb14b81f442eb57a38755829270ff2c4315:20
ralonsohok, I need to review this patch too (I remember it)15:21
ralonsohI think this is working by accident15:22
opendevreviewPhilipp Dreesens proposed openstack/nova-specs master: Add spec for bidirectional RPC liveness handshake  https://review.opendev.org/c/openstack/nova-specs/+/98438415:22
ralonsoh--> The network-vif-plugged during live migration was sent by the SOURCE agent calling update_device_up (for the existing binding that's still active on the source), NOT by the destination agent.15:22
sean-k-mooneythat shoudl not happen becuse we set migrating_to15:23
sean-k-mooneyin the port profiel15:23
ralonsohI know, I changed this code15:23
sean-k-mooneyand only the host named there is ment to send it with yes you updated that code15:24
ralonsohsean-k-mooney, thanks a lot, I need to recheck the recent changes in the OVS agent15:25
ralonsoh(done by me, trying to fix other issues)15:26
sean-k-mooneyas an aside "``os_vif_delegation=True`` and ``create_port=True``" will mena the tap is create before we add the ovs port15:32
sean-k-mooneyso in that case we wont see -1 for kernel ovs but we still will for dpdk15:32
sean-k-mooneyos we shoudl not have divergent code paths if we can aovid it and just alwasy supprot -115:32
ralonsohbut isn't supporting ofport=-1 something wrong?15:38
ralonsohwith  ``create_port=True``, when the port is created? who creates it?15:39
ralonsohI need to leave now, I'll check this chat later. I'll also review the current OVS agent code15:40
sean-k-mooneynot in general. with create_port=True its not required15:42
sean-k-mooneyfor kernel ovs15:42
sean-k-mooneythis is not the first time we have had to revert skipping prot wot ofprot=-1 https://github.com/openstack/neutron/commit/f1fa9529ee6753e2a3f47e4a6eb0286118520d2515:44
opendevreviewGoutham Pacha Ravi proposed openstack/nova master: Allow rebuild for instances with virtiofs shares  https://review.opendev.org/c/openstack/nova/+/99904821:44
opendevreviewGhanshyam Maan proposed openstack/nova master: Task tracking mechanism for graceful shutdown  https://review.opendev.org/c/openstack/nova/+/99629922:12
gmaanmelwitt: thanks for review, ^^ updated the wrapper cache feedback. To make revert_resize better in shutdown, we need more work on some RPC redesign or so. I added the details in gerrit and reported a separate bug for tracking it for graceful shutdown part-3 or something to fix as a separate bug https://bugs.launchpad.net/nova/+bug/216209022:21
gmaanwe need to make revert_resize more sync between source and dest compute like live migration or other operations which involves two compute22:22
melwittgmaan: ack thanks22:45
opendevreviewGhanshyam Maan proposed openstack/nova master: Call manager graceful_shutdown while RPC server waits  https://review.opendev.org/c/openstack/nova/+/99658322:46

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!