June 2022 - openstack-discuss - lists.openstack.org

[kolla-ansible][xena] Facing network issue when providing SR-IOV support
by Harshit Mishra 22 Jun '22

22 Jun '22

Hi! I have deployed Openstack Xena using kolla-ansible (v13.0.1) on a multi-node setup (1 controller+network, multiple computes). On this cluster I would like to have support for normal, as well as direct (SR-IOV) vNIC types. I have done all the pre-requirement like configure VF on SR-IOV network interface. In the current deployment, I have created two physnets, one for flat network (called physnet1 on br-ex), and one for SR-IOV (called sriovnet1 on Intel 10G 2P X520 card). I am creating one network of VXLAN type for normal vNIC, and one of VLAN type on sriovnet1 (called sriov_network) for direct vNIC. Mapping of PF with provider in sriov-agent.ini: # cat /etc/kolla/neutron-sriov-agent/sriov_agent.ini [sriov_nic] physical_device_mappings = sriovnet1:enp8s0f1 exclude_devices = [securitygroup] firewall_driver = neutron.agent.firewall.NoopFirewallDriver ---------------------------------------------- ml2_conf.ini on controller node: # cat /etc/kolla/neutron-server/ml2_conf.ini [ml2] type_drivers = flat,vlan,vxlan tenant_network_types = vxlan,vlan,flat mechanism_drivers = openvswitch,l2population,sriovnicswitch extension_drivers = port_security [ml2_type_vlan] network_vlan_ranges = physnet1,sriovnet1:1000:1009 [ml2_type_flat] flat_networks = physnet1,sriovnet1 [ml2_sriov] agent_required = False supported_pci_vendor_devs = 8086:10ed [ml2_type_vxlan] vni_ranges = 1:1000 ---------------------------------------------- # grep -nr enabled_filters /etc/kolla/nova-scheduler/ /etc/kolla/nova-scheduler/nova.conf:13:enabled_filters = ComputeFilter,AvailabilityZoneFilter,ComputeCapabilitiesFilter,ImagePropertiesFilter,ServerGroupAntiAffinityFilter,ServerGroupAffinityFilter,SameHostFilter,DifferentHostFilter,AggregateInstanceExtraSpecsFilter,PciPassthroughFilter ---------------------------------------------- # grep passthrough /etc/kolla/nova-compute/nova.conf passthrough_whitelist = [{"physical_network": "sriovnet1", "devname": "enp8s0f1"}] The interfaces on VMs from normal vNICs are able to communicate inter-compute. The same is not working for the direct vNICs belonging to the sriov_network. I have tried multiple config changes, but none seem to work. Please help in solving this issue, or suggest a different way of achieving this if this is wrong or not optimal. Thanks! Regards, Harshit Mishra

1 0

[Dev] [Skyline] Boot from Image directly
by Oezcan Isik 22 Jun '22

22 Jun '22

Hi, Will Skyline support boot from image directly during instance creation. Instead of choosing a system disk (volume) every time. Kind Regards, -- Ozcan Isik Software Engineer STAG CERN IT

2 1

[TripleO] no more patches for tripleo stable/victoria please (going EOL)
by Marios Andreou 21 Jun '22

21 Jun '22

Hello o/ as proposed at [1][2] we are moving ahead with transitioning stable/victoria tripleo repos to End Of Life. In order to do this we need to have no open patches against stable/victoria. So please stop posting new patches for tripleo stable/victoria and if you have open patches please merge or abandon these by next Tuesday 21st June (for reference, list of repos with currently open patches at [3]). No problem if someone needs a few more days we can wait just please reach out here or on irc. If there are no such requests then I will abandon all open reviews on 21st and respin [2] to pick up any new commits between now and then. thanks for your help, marios [1] http://lists.openstack.org/pipermail/openstack-discuss/2022-June/028829.html [2] https://review.opendev.org/c/openstack/releases/+/845148 [3] (list of repos with open reviews): * https://review.opendev.org/q/project:openstack%252Fos-net-config+status:ope… * https://review.opendev.org/q/project:openstack%252Fpython-tripleoclient+sta… * https://review.opendev.org/q/project:openstack%252Ftripleo-ansible+status:o… * https://review.opendev.org/q/project:openstack%252Ftripleo-common+status:op… * https://review.opendev.org/q/project:openstack%252Ftripleo-heat-templates+s… * https://review.opendev.org/q/project:openstack%252Ftripleo-image-elements+s… * https://review.opendev.org/q/project:openstack%252Ftripleo-puppet-elements+… * https://review.opendev.org/q/project:openstack%252Ftripleo-validations+stat…

4 8

Propose to add Takashi Kajinami as heat core reviewer
by Rico Lin 21 Jun '22

21 Jun '22

Hi all I want to suggest adding Takashi Kajinami <tkajinam(a)redhat.com> as heat core reviewer. Who help provide valid reviews and fixes for issues on heat. Will be great to have Takashi Kajinami join heat core reviewer team and help. *Rico Lin*

4 4

Re: Regarding Floating IP is existing Setup
by Eugen Block 21 Jun '22

21 Jun '22

Could this be a race condition? I saw something similar, too, when a user tried to attach floating IPs via terraform or ansible (I'm not quite sure) at the same time to two different VMs. We could trace it in the Logs that the port was first attached to the first instance and then failed for the other. I don't recall if they use a workaround at the moment or if just works now, will need to check with them. If this is a race condition can you do anything about it in Heat? Like some sort of delay? It could be a bug, though. Zitat von Adivya Singh <adivya1.singh(a)gmail.com>: > Hi Eugen, > > These are orchestration Heat templates, if there is a Single FLoating IP > creation using this templates it works fine, But if there are multiple > floating IP attachment it shows there is a issue on 2 to 3 floating IP > attached to the instance > > Regards > Adivya Singh > > On Tue, Jun 21, 2022 at 5:59 PM Eugen Block <eblock(a)nde.ag> wrote: > >> Do you have ha routers enabled? >> >> > Most of the template having issue , Have all instance having FLoating >> IP, a >> > Stack with a single floating IP have chance of issue very less >> >> I don't quite understand that sentence, could you please clarify? >> >> Zitat von Adivya Singh <adivya1.singh(a)gmail.com>: >> >> > hi Eugen, >> > >> > The current setup is 3 controller nodes, The Load is not much on each >> > controller and the number of DHCP agent is always set to 2 as per the >> > standard in the neutron.conf, The L3 agent seems to be stables as other >> > router namespace works fine under it, Only few router Namespace get >> > affected under the agent. >> > >> > Most of the template having issue , Have all instance having FLoating >> IP, a >> > Stack with a single floating IP have chance of issue very less >> > >> > Regards >> > Adivya Singh >> > >> > On Tue, Jun 21, 2022 at 1:18 PM Eugen Block <eblock(a)nde.ag> wrote: >> > >> >> Hi, >> >> >> >> this sounds very familiar to me, I had to deal with something similar >> >> a couple of times in a heavily used cluster with 2 control nodes. What >> >> does your setup look like, is it a HA setup? I would start checking >> >> the DHCP and L3 agents. After increasing dhcp_agents_per_network to 2 >> >> in neutron.conf and restarting the services this didn't occur again >> >> (yet). This would impact floating IPs as well, sometimes I had to >> >> disable and enable the affected router(s). If you only have one >> >> control node a different approach is necessary. Do you see a high load >> >> on the control node? >> >> >> >> >> >> Zitat von Adivya Singh <adivya1.singh(a)gmail.com>: >> >> >> >> > hi Team, >> >> > >> >> > We got a issue in Xena release, where we set the environment in Ubuntu >> >> > Platform, But later we get some issues in Floating IP not reachable. >> >> > >> >> > In a Network node, not all router namespace are Impacted and only few >> of >> >> > them get affected, So we can not comment Network node has a issue. >> >> > >> >> > The L3 agent where the Router is tied up, Worked just fine, as other >> >> > routers work Fine. >> >> > >> >> > and the one having issue in Floating IP, if i unassigned and assigned >> it >> >> > starts working most of the time. >> >> > >> >> > Any thoughts on this >> >> > >> >> > Regards >> >> > Adivya Singh >> >> >> >> >> >> >> >> >> >> >> >> >> >> >>

1 0

[Tripleo] - IPv6 Wallaby Undercloud Installation failure
by Anirudh Gupta 21 Jun '22

21 Jun '22

Hi Team, I am trying to deploy Openstack Wallaby Undercloud on IPv6, but facing the below error: 2022-06-14 05:01:23.213708 | 52540083-cfa2-3f20-e9dc-00000000286f | TASK | Manage container systemd services and cleanup old systemd healthchecks for /var/lib/tripleo-config/container-startup-config/step_4 2022-06-14 05:03:22.912816 | 52540083-cfa2-3f20-e9dc-00000000286f | FATAL | Manage container systemd services and cleanup old systemd healthchecks for /var/lib/tripleo-config/container-startup-config/step_4 | undercloud | error={"changed": false, "msg": "Service ironic_pxe_tftp has not started yet"} 2022-06-14 05:03:22.914400 | 52540083-cfa2-3f20-e9dc-00000000286f | TIMING | tripleo_container_manage : Manage container systemd Sample Undercloud.conf is as follows: [DEFAULT] clean_nodes = true cleanup = false container_cli = podman container_healthcheck_disabled = true container_images_file = /home/stack/containers-prepare-parameter.yaml deployment_user = stack enable_ironic = true enable_ironic_inspector = true enable_neutron = true enable_routed_networks = false generate_service_certificate = false ipv6_address_mode = dhcpv6-stateful ipxe_enabled = true local_interface = enp8s0 local_ip = aaaa:aaaa:aaaa::1/64 subnets = ctlplane-subnet undercloud_admin_host = aaaa:aaaa:aaaa::1 undercloud_hostname = undercloud.com undercloud_ntp_servers = 30.30.30.3 undercloud_public_host = aaaa:aaaa:aaaa::1 undercloud_timezone = UTC [ctlplane-subnet] cidr = aaaa:aaaa:aaaa::/64 dhcp_end = aaaa:aaaa:aaaa::f dhcp_start = aaaa:aaaa:aaaa::a gateway = aaaa:aaaa:aaaa::1 inspection_iprange = aaaa:aaaa:aaaa::3,aaaa:aaaa:aaaa::9 Can someone please help in this regard. Anirudh Gupta

7 21

[nova] custom metric for weigh scheduling
by Arnaud Morin 21 Jun '22

21 Jun '22

Hey nova team, Some context: I want nova-scheduler to favor instances on some computes based on some hypervisor metrics/traits. E.G. I have 2 hypervisors hosting few instances. One of them is having an old kernel, that I will upgrade some day. The second is already having the correct target kernel. I want the scheduler to favor booting on the second host (good kernel). How can I achieve that without disabling the bad_kernel hypervisor? I was thinking about something in the weight scheduler. I was first thinking about using placement traits (having a custom traits) and reading traits from a weight scheduler. But it seems not doable out of the box. And it also seems not a good idea at all to call placement API from weight scheduler. Maybe I am missing something here because I was expecting to have the traits given by placement up to the nova scheduler. Second shoot, I found the metrics weight [1] which rely on the compute monitors. It seems less hacky, but I will still have to write custom monitor to retrieve such metrics. Am I missing an option here? Is there anyone doing such thing is the community? [1] https://github.com/openstack/nova/blob/master/nova/scheduler/weights/metric… [2] https://github.com/openstack/nova/tree/master/nova/compute/monitors Cheers, Arnaud.

1 0

Re: Few Open Questions: Openstack vs Proprietary cloud stacks
by Lazuardi Nasution 21 Jun '22

21 Jun '22

> > > Hi Krish, Please find inline comment based my experience on banking implementation. Best regards. > Date: Tue, 21 Jun 2022 12:29:01 +0530 > From: KK CHN <kkchn.in(a)gmail.com> > To: openstack-discuss(a)lists.openstack.org > Subject: Few Open Questions: Openstack vs Proprietary cloud stacks > Message-ID: > <CAKgGyB8L-tQKR49Ru4tqCzox5C5o57sJZ+LbfHUpxBu== > ecxyQ(a)mail.gmail.com> > Content-Type: text/plain; charset="utf-8" > > List, > > Very recently I came across a discussion with a banking corporation who > are heavy users of VMware in all their Data Centre operations .. > > They asked me a lot of questions in comparison with openstack vs other > proprietary vendors (such as VMware. ). > > > As I am promoting OpenStack I need to clarify these questions raised by > the organization. > > I request your inputs for these queries . > > 1. Does OpenStack support Hyper Converged Infrastructure ? > Or any HCI vendors already with OpenStack solutions? Please share the > URL/link of case studies of OpenStack HCI. > Any documentation/Reference to install openstack as HCI solution with > supported H/W OEM details . I can try out ... > Yes, I have experienced with HCI by using OpenStack and Ceph. > 2. Is there any performance benchmark for openstack to compare with other > such cloud stacks ( URL/reference links kindly share ) > I have done some performance test by using CBTool which is the corr of SPEC Cloud IaaS benchmark. So, so the result can be compared with published SPEC Cloud IaaS result as long it has same benchmark parameters. > 3. What is the market share of OpenStack all over the world ? ( If any > statistics kindly share.) > Any survey reports available how many large and medium organizations are > using Openstack distributions for private/public cloud use in production ? > Maybe you can refer to OpenStack surveys. It seem that on any forms (including customized), OpenStack os popular for public cloud. > 4. Does OpenStack new versions support VDI(Virtual Desktop Infrastructure) > solutions out of the box.? pls give a reference link / url for the same. > Actually, VDI consists managed desktop VMs. I haven't use it but maybe Guacamole can be used for VDI on OpenStack. > 5. What is the alternative for VROPS and VRA (including self service > capabilities, Blueprint/Template based deployment) in OpenStack and any > comparative study on the capabilities available. ( Please share inputs if > any to refer. ) > OpenStack includes self service capability. Yes, it has template based deployment like Heat and catalog based deployment like Murano. > 6. What is the alternative of VMWARE Network Insights and VMWARE > CodeStream and Vrealize Loginsights in OpenStack ? > I use other open source products for monitoring and log analysis. Somethink like Prometheus, Loki, etc. > 7. Any major capability available in VMware cloud stack for which > alternative is not available in OpenStack out of the box ? > I have no idea about this. But it is hard to have apple to apple comparison. > 8. Tools to maintain multiple OpenStack Cluster, any concept of > Supervisory cluster available There are some cloud management products which are compatible with OpenStack. But for me, Ansible and/or BASH scripting are good enough. > 9. Who all are the large organizations who deployed OpenStack as their > alternative over proprietary solutions(eg: VMware cloud and virtualization > products ..) > I have no idea with this. > 10. Does OpenInfra Foundation also provide direct support For OpenStack > Deployment ? If yes Kindly share the link / url where to contact > /approach for paid support ? in case the clients need paid support ? > It seem OpenInfra doesn't, but OpenInfra member do. CMIIW. > > 11. Regarding security of cloud environments /infrastructure, what are the > USPs of OpenStack cloud security over other competing cloud middleware > providers ? and the vulnerability handling model/mechanism in production ? > On OpenStack, every tenants can be fully isolated. Anyway, some security products may have another approach to enhance the security. Floating IP concept is good for hiding non service (externally) instance. > Thanks in advance for your invaluable inputs. > > Each and every input from each member is valuable for me to showcase as > solid proofs to defend and promote the openstack over proprietary market > leaders. > > Krish >

1 0

[cinder] cinder backup benji driver.
by Jesper Schmitz Mouridsen 21 Jun '22

21 Jun '22

Hi, I have been looking into using benji-backup[1] in a cinder-backup driver[2]. I have not tested it thoroughly enough yet, but both in devstack and in a multinode setup.(focal victoria). On devstack I get a newer sqlalchemy that satisfies Benji's requirements list. I have not found issues with downgrading Benji to use an earlier version of sqlalchemy found in canonical focal repos. But that also needs some testing. I asked the author of Benji why he chose a version as high as he did. So just a little info about this work. Any comments on the idea? Thanks. Jesper Schmitz Mouridsen. [1] https://benji-backup.me/ [2] https://github.com/jsm222/cinder-backup-benji-driver

1 0

[neutron] CI meeting 21.06 cancelled
by Slawek Kaplonski 21 Jun '22

21 Jun '22

Hi, I will be on PTO this Tuesday and I will not be able to run Neutron CI meeting. We don't have anything really urgent what can't wait one week more so lets cancel this week's meeting. See You on the Neutron CI meeting next week. -- Slawek Kaplonski Principal Software Engineer Red Hat

1 0