- openstack-discuss - lists.openstack.org

Re: [openstack-dev] [tripleo] Zuul Queue backlogs and resource usage
by Bogdan Dobrelya 26 Nov '18

26 Nov '18

Here is a related bug [1] and implementation [1] for that. PTAL folks! [0] https://bugs.launchpad.net/tripleo/+bug/1804822 [1] https://review.openstack.org/#/q/topic:base-container-reduction > Let's also think of removing puppet-tripleo from the base container. > It really brings the world-in (and yum updates in CI!) each job and each > container! > So if we did so, we should then either install puppet-tripleo and co on > the host and bind-mount it for the docker-puppet deployment task steps > (bad idea IMO), OR use the magical --volumes-from <a-side-car-container> > option to mount volumes from some "puppet-config" sidecar container > inside each of the containers being launched by docker-puppet tooling. On Wed, Oct 31, 2018 at 11:16 AM Harald Jensås <hjensas at redhat.com> wrote: > We add this to all images: > > https://github.com/openstack/tripleo-common/blob/d35af75b0d8c4683a677660646… > > /bin/sh -c yum -y install iproute iscsi-initiator-utils lvm2 python > socat sudo which openstack-tripleo-common-container-base rsync cronie > crudini openstack-selinux ansible python-shade puppet-tripleo python2- > kubernetes && yum clean all && rm -rf /var/cache/yum 276 MB > > Is the additional 276 MB reasonable here? > openstack-selinux <- This package run relabling, does that kind of > touching the filesystem impact the size due to docker layers? > > Also: python2-kubernetes is a fairly large package (18007990) do we use > that in every image? I don't see any tripleo related repos importing > from that when searching on Hound? The original commit message[1] > adding it states it is for future convenience. > > On my undercloud we have 101 images, if we are downloading every 18 MB > per image thats almost 1.8 GB for a package we don't use? (I hope it's > not like this? With docker layers, we only download that 276 MB > transaction once? Or?) > > > [1] https://review.openstack.org/527927 -- Best regards, Bogdan Dobrelya, Irc #bogdando

1 0

[openstack-dev] [nova][placement] Please help to review XenServer vgpu related patches
by Naichuan Sun 26 Nov '18

26 Nov '18

Hi, Sylvain, Jay, Eric and Matt, I saw the n-rp and reshaper patches in upstream have almost finished. Could you help to review XenServer vGPU related patches when you have the time? https://review.openstack.org/#/c/520313/ https://review.openstack.org/#/c/521041/ https://review.openstack.org/#/c/521717/ Thank you very much. BR. Naichuan Sun __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

1 0

[neutron] LP tag for python 3
by Takashi Yamamoto 26 Nov '18

26 Nov '18

hi, currently we seem to have an official tag "py34" on the LP. and "py35" is documented in the contributor guide. https://docs.openstack.org/neutron/latest/contributor/policies/bugs.html#py… i'd like to propose to replace them with "py3" mechanically.

2 1

[Openstack] [OpenStack][Cinder]How config cinder.conf [nova] part for InstanceLocalityFilter
by Soheil Pourbafrani 26 Nov '18

26 Nov '18

Hi, I want to enable InstanceLocalityFiletr for the Cinder. I set scheduler_default_filters = AvailabilityZoneFilter,CapacityFilter,CapabilitiesFilter,InstanceLocality But the document said we should specify an account with privileged rights for Nova in Cinder configuration (configure a keystone authentication plugin in the [nova] section), but I can't find an example for such configuration. Can anyone give an example? _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack(a)lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

1 0

[Openstack] [OpenStack][Cinder] Config Volumes to Store on Compute nodes
by Soheil Pourbafrani 26 Nov '18

26 Nov '18

Hi, I want to configure cinder to store Volumes on the Compute nodes. I know I should use the InstanceLocalityFilter. Suppose we have one Controller and two Compute node, should I install the Cinder on each node (Controller and Compute)? I will be appreciated if you suggest me any instruction about this. Thanks! _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack(a)lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

1 0

[Openstack] [openstack][neutron]Iptables snat not work when securitygroup is on
by 陈炤 25 Nov '18

25 Nov '18

Hi, I am building an openstack all-in-one environment in a CentOS7.4 machine. For some reason I have only one network interface(eth0) and one ip address, so I created a linux bridge(br0), and forwarded datas to eth0 using iptables command: iptables -t nat -A POSTROUTING -s {bridge virtual ip} -j SNAT --to {eth0 ip} But it seems not work. When I ping to 8.8.8.8 from br0 and run tcpdump, I can see that datas can be forwared to eth0 and be sent to 8.8.8.8, but when datas are sent back to eth0, they can not be forwarded to br0. Ip forwarding, net.bridge.bridge-nf-call-iptables and net.bridge.bridge-nf-call-ip6tablesare set to 1. If I close security group by setting securitygroup = false, this rule works fine, but if I use iptables -F instead, the rule is not work. Does the securitygroup have a magic to trap iptables? BR Don _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack(a)lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

2 1

[openstack-dev] [kolla] Berlin summit resume
by Eduardo Gonzalez 25 Nov '18

25 Nov '18

Hi kollagues, During the Berlin Summit kolla team had a few talks and forum discussions, as well as other cross-project related topics [0] First session was ``Kolla project onboarding``, the room was full of people interested in contribute to kolla, many of them already using kolla in production environments whiling to make upstream some work they've done downstream. I can say this talk was a total success and we hope to see many new faces during this release putting features and bug fixes into kolla. Slides of the session at [1] Second session was ``Kolla project update``, was a brief resume of what work has been done during rocky release and some items will be implemented in the future. Number of attendees to this session was massive, no more people could enter the room. Slides at [2] Then forum sessions.. First one was ``Kolla user feedback``, many users came over the room. We've notice a big increase in production deployments and some PoC migrating to production soon, many of those environments are huge. Overall the impressions was that kolla is great and don't have any big issue or requirement, ``it works great`` became a common phrase to listen. Here's a resume of the user feedback needs [3] - Improve operational usage for add, remove, change and stop/start nodes and services. - Database backup and recovery - Lack of documentation is the bigger request, users need to read the code to know how to configure other than core/default services - Multi cells_v2 - New services request, cyborg, masakari and tricircle were the most requested - SElinux enabled - More SDN services such as Contrail and calico - Possibility to include user's ansible tasks during deploy as well as support custom config.json - HTTPS for internal networks Second one was about ``kolla for the edge``, we've meet with Edge computing group and others interested in edge deployments to identify what's missing in kolla and where we can help. Things we've identified are: - Kolla seems good at how the service split can be done, tweaking inventory file and config values can deploy independent environments easily. - Missing keystone federation - Glance cache support is not a hard requirement but improves efficiency (already merged) - Multi cells v2 - Multi storage per edge/far-edge - A documentation or architecture reference would be nice to have. Last one was ``kolla for NFV``, few people came over to discuss about NUMA, GPU, SRIOV. Nothing noticiable from this session, mainly was support DPDK for CentOS/RHEL,OracleLinux and few service addition covered by previous discussions. [0] https://etherpad.openstack.org/p/kolla-stein-summit [1] https://es.slideshare.net/EduardoGonzalezGutie/kolla-project-onboarding-ope… [2] https://es.slideshare.net/EduardoGonzalezGutie/openstack-kolla-project-upda… [3] https://etherpad.openstack.org/p/berlin-2018-kolla-user-feedback [4] https://etherpad.openstack.org/p/berlin-2018-kolla-edge [5] https://etherpad.openstack.org/p/berlin-2018-kolla-nfv __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

3 2

[Openstack] [OpenStack][Glance]Getting instance snapshot result in size 0 byte
by Soheil Pourbafrani 24 Nov '18

24 Nov '18

Hi, I have many instances running on OpenStack and I wanted to export them. So I create a snapshot of an instance and it results in a new record in images in the format of Qcow2 and size of 0 bytes! It just created a volume snapshot of the instance, too. I tried with both command line and horizon but the same results! How can I export instances in the correct way? _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack(a)lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

2 1

[Openstack] [OpenStack][Keystone] the controller is node listening on port 5000
by Soheil Pourbafrani 24 Nov '18

24 Nov '18

Hi, I installed keystone step by step according to the OpenStack Doc <https://docs.openstack.org/keystone/rocky/install/keystone-install-rdo.html> but the http server is not listening on the port 5000! The only difference is my hostname (address) is "controller.a.b" instead of "controller". In verifying part of the document I could do all (like creating a user, project, ....) but after installing glance, I found out the Glance API is not running properly and it's because of Keystone is not listening on port 5000! Here are AUTH variables: export OS_PROJECT_DOMAIN_NAME=Default export OS_USER_DOMAIN_NAME=Default export OS_PROJECT_NAME=admin export OS_USERNAME=admin export OS_PASSWORD=XXX export OS_AUTH_URL=http://controller.a.b:5000/v3 export OS_IDENTITY_API_VERSION=3 export OS_IMAGE_API_VERSION=2 When I try "glance image-list", I got: "http://controller.a.b:9292/v2/images?limit=20&sort_key=name&sort_dir=asc: Unable to establish connection" Here is TCP listening ports: tcp 0 0 127.0.0.1:25 0.0.0.0:* LISTEN 1614/master tcp 0 0 0.0.0.0:9191 0.0.0.0:* LISTEN 961/python2 tcp 0 0 0.0.0.0:25672 0.0.0.0:* LISTEN 957/beam.smp tcp 0 0 192.168.0.31:3306 0.0.0.0:* LISTEN 1622/mysqld tcp 0 0 192.168.0.31:2379 0.0.0.0:* LISTEN 963/etcd tcp 0 0 192.168.0.31:11211 0.0.0.0:* LISTEN 952/memcached tcp 0 0 127.0.0.1:11211 0.0.0.0:* LISTEN 952/memcached tcp 0 0 192.168.0.31:2380 0.0.0.0:* LISTEN 963/etcd tcp 0 0 0.0.0.0:4369 0.0.0.0:* LISTEN 1/systemd tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 953/sshd tcp6 0 0 ::1:25 :::* LISTEN 1614/master tcp6 0 0 :::5000 :::* LISTEN 960/httpd tcp6 0 0 :::5672 :::* LISTEN 957/beam.smp tcp6 0 0 ::1:11211 :::* LISTEN 952/memcached tcp6 0 0 :::80 :::* LISTEN 960/httpd tcp6 0 0 :::22 :::* LISTEN 953/sshd As you can see httpd is listening on port 5000 IPV6, but it's not active for IPV4! Here is keystone.log: 2018-11-24 15:55:59.685 17325 INFO keystone.common.wsgi [req-9fedd505-5d98-4c12-a6ad-967f28c766e7 - - - - -] POST http://controller.a.b:5000/v3/auth/tokens 2018-11-24 15:56:00.275 17327 INFO keystone.common.wsgi [req-ca93da87-1d9f-4611-835c-d6972054b8e1 - - - - -] POST http://controller.a.b:5000/v3/auth/tokens 2018-11-24 15:56:00.929 17328 INFO keystone.common.wsgi [req-cf8c6af8-342d-4b14-9826-ba27e5f2897a 1b96f6b67f08495092e98a9b60476152 b33579097090499499625154e92724ee - default default] GET http://controller.a.b:5000/v3/domains/default 2018-11-24 15:56:14.753 17325 INFO keystone.common.wsgi [req-cca0e4a5-0ef9-40d8-a2f4-56c1f0acafcc 1b96f6b67f08495092e98a9b60476152 b33579097090499499625154e92724ee - default default] POST http://controller.a.b:5000/v3/users 2018-11-24 15:56:14.773 17325 WARNING py.warnings [req-cca0e4a5-0ef9-40d8-a2f4-56c1f0acafcc 1b96f6b67f08495092e98a9b60476152 b33579097090499499625154e92724ee - default default] /usr/lib/python2.7/site-packages/oslo_policy/policy.py:896: UserWarning: Policy identity:create_user failed scope check. The token used to make the request was project scoped but the policy requires ['system'] scope. This behavior may change in the future where using the intended scope is required warnings.warn(msg) Another difference between my system and document is that I use static IP and here are interface settings: TYPE="Ethernet" PROXY_METHOD="none" BROWSER_ONLY="no" BOOTPROTO="none" DEFROUTE="yes" IPV4_FAILURE_FATAL="no" IPV6INIT="yes" IPV6_AUTOCONF="yes" IPV6_DEFROUTE="yes" IPV6_FAILURE_FATAL="no" IPV6_ADDR_GEN_MODE="stable-privacy" NAME="enp2s0" UUID="62aa5ce8-64da-41b4-abd3-293fd43584fa" DEVICE="enp2s0" ONBOOT="yes" IPADDR="192.168.0.X" PREFIX="24" GATEWAY="192.168.0.1" DNS1="192.168.0.1" DNS2="8.8.8.8" IPV6_PRIVACY="no" What can be the cause? _______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack(a)lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

1 0

[oslo] [config] Using environment variables for config options
by Chris Dent 24 Nov '18

24 Nov '18

I don't recall seeing any specific announcement for this, and I think it is a pretty useful feature, so I thought I'd write about some functionality that I added to oslo.config and was released in 6.7.0. Especially since a lot of people I've mentioned this to thought we already had it. The new version is able to satisfy any registered config option by looking in the process environment for a variable with a predictable name and using its value if set. Options are satisfied in this order: * command line args * environment * config file(s) * everything else (using the new(ish) "sources" concept) In the past it was possible to get config from the environment if the option had been pre-defined as using something in the environment as its default. This new way is much more flexible, and in my experience ideal for container-based environments where you would like an immutable container but want small differences between multiple instances of the same container to be easy to control. Or if you don't need much config at all, you can avoid using a config file entirely. I've been using this with placement, where it's possible to set up a testable placement service with just two environment variables: OS_PLACEMENT_DATABASE__CONNECTION=mysql+pymysql://root:secret@127.0.0.1/placement?charset=utf8 OS_API__AUTH_STRATEGY=noauth2 This shows the pattern of the environment variable name: OS_<config group>__<config opt> There are some docs at https://docs.openstack.org/oslo.config/latest/reference/drivers.html#module… I hope this proves as useful for other people as it has for me. -- Chris Dent ٩◔̯◔۶ https://anticdent.org/ freenode: cdent tw: @anticdent

3 3