April 2022 - openstack-discuss - lists.openstack.org

Problems running Watcher on DevStack
by Christian Rohmann 05 Apr '22

05 Apr '22

Hello there, I am trying to setup devstack which is running Watcher. But I am running always into problems with setup not finishing and throwing errors (see below). To clarify: I am trying to setup the devstack with multiple nodes (three nodes to be exactly, one control/compute and two compute) following the instructions at https://docs.openstack.org/watcher/latest/contributor/devstack.html which leads to this configuration control / first compute node: > [[local|localrc]] > ADMIN_PASSWORD="<password>" > DATABASE_PASSWORD=$ADMIN_PASSWORD > RABBIT_PASSWORD=$ADMIN_PASSWORD > SERVICE_PASSWORD=$ADMIN_PASSWORD > SERVICE_HOST="<ip of the control node>" > HOST_IP="<ip of host>" > FIXED_RANGE="10.4.128.0/20" > FLOATING_RANGE="subnet range" > > DEVSTACK_RELEASE="stable/yoga" > > LOGFILE=/opt/stack/logs/stack.sh.log > LOGDAYS=7 > LOG_COLOR=False > > enable_plugin watcher https://opendev.org/openstack/watcher > $DEVSTACK_RELEASE > > enable_plugin watcher-dashboard > https://opendev.org/openstack/watcher-dashboard $DEVSTACK_RELEASE > > enable_plugin ceilometer https://opendev.org/openstack/ceilometer.git > $DEVSTACK_RELEASE > CEILOMETER_BACKEND=gnocchi > > enable_service ceilometer-api > enable_service ceilometer-acompute > > enable_plugin gnocchi https://github.com/gnocchixyz/gnocchi > > # OpenStack Telemetry (Ceilometer) Alarming > # enable_plugin aodh https://opendev.org/openstack/aodh $DEVSTACK_RELEASE > > # I did not use the panko project which is mentioned in the > documentation because it is deprecated > > [[post-config|$NOVA_CONF]] > [DEFAULT] > compute_monitors=cpu.virt_driver > [scheduler] > discover_hosts_in_cells_interval=2 additional compute node(s): > [[local|localrc]] > HOST_IP="<host ip>" > FIXED_RANGE=10.4.128.0/20 > FLOATING_RANGE="<floating range" > ADMIN_PASSWORD="<password>" > DATABASE_PASSWORD=$ADMIN_PASSWORD > RABBIT_PASSWORD=$ADMIN_PASSWORD > SERVICE_PASSWORD=$ADMIN_PASSWORD > DATABASE_TYPE=mysql > SERVICE_HOST="<ip of control node>" > MYSQL_HOST=$SERVICE_HOST > RABBIT_HOST=$SERVICE_HOST > GLANCE_HOSTPORT=$SERVICE_HOST:9292 > Q_HOST=$SERVICE_HOST > OVN_SB_REMOTE=tcp:$SERVICE_HOST:6642 > disable_all_services > ENABLED_SERVICES=n-cpu,placement-client,ovn-controller,q-ovn-metadata-agent > NOVA_VNC_ENABLED=True > NOVNCPROXY_URL="http://$SERVICE_HOST:6080/vnc_lite.html" > VNCSERVER_LISTEN=$HOST_IP > VNCSERVER_PROXYCLIENT_ADDRESS=$VNCSERVER_LISTEN > > DEVSTACK_RELEASE="stable/yoga" > > LOGFILE=/opt/stack/logs/stack.sh.log > LOGDAYS=7 > LOG_COLOR=False > > enable_plugin ceilometer https://opendev.org/openstack/ceilometer > $DEVSTACK_RELEASE > disable_service ceilometer-acentral > disable_service ceilometer-collector > disable_service ceilometer-api > > [[post-config|$NOVA_CONF]] > [DEFAULT] > compute_monitors=cpu.virt_driver In addition I have tried to use the configuration that can be found in the repository of the watcher-project at https://opendev.org/openstack/watcher/src/branch/master/devstack. The error is always the same with setup.sh ending in: > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:244 > : cp /opt/stack/ceilometer/etc/ceilometer/polling_all.yaml > /etc/ceilometer/polling.yaml > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:246 > : cp > /opt/stack/ceilometer/ceilometer/pipeline/data/event_definitions.yaml > /opt/stack/ceilometer/ceilometer/pipeline/data/event_pipeline.yaml > /opt/stack/ceilometer/ceilometer/pipeline/data/pipeline.yaml > /etc/ceilometer > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:248 > : '[' '' ']' > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:251 > : '[' False == True ']' > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:259 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > auth_type password > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:260 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > user_domain_id default > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:261 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > project_domain_id default > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:262 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > project_name service > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:263 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > username ceilometer > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:264 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > password <password> > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:265 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > region_name RegionOne > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:266 > : iniset /etc/ceilometer/ceilometer.conf service_credentials > auth_url http://192.168.0.100/identity > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:268 > : configure_auth_token_middleware /etc/ceilometer/ceilometer.conf > ceilometer /var/cache/ceilometer > ++ lib/keystone:configure_auth_token_middleware:467 : echo 'WARNING: > configure_auth_token_middleware is deprecated, use > configure_keystone_authtoken_middleware instead' > WARNING: configure_auth_token_middleware is deprecated, use > configure_keystone_authtoken_middleware instead > ++ lib/keystone:configure_auth_token_middleware:468 : > configure_keystone_authtoken_middleware > /etc/ceilometer/ceilometer.conf ceilometer > ++ lib/keystone:configure_keystone_authtoken_middleware:447 : local > conf_file=/etc/ceilometer/ceilometer.conf > ++ lib/keystone:configure_keystone_authtoken_middleware:448 : local > admin_user=ceilometer > ++ lib/keystone:configure_keystone_authtoken_middleware:449 : local > section=keystone_authtoken > ++ lib/keystone:configure_keystone_authtoken_middleware:451 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken auth_type password > ++ lib/keystone:configure_keystone_authtoken_middleware:452 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken interface public > ++ lib/keystone:configure_keystone_authtoken_middleware:453 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken auth_url > http://192.168.0.100/identity > ++ lib/keystone:configure_keystone_authtoken_middleware:454 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken username ceilometer > ++ lib/keystone:configure_keystone_authtoken_middleware:455 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken password <password> > ++ lib/keystone:configure_keystone_authtoken_middleware:456 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken user_domain_name > Default > ++ lib/keystone:configure_keystone_authtoken_middleware:457 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken project_name service > ++ lib/keystone:configure_keystone_authtoken_middleware:458 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken project_domain_name > Default > ++ lib/keystone:configure_keystone_authtoken_middleware:460 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken cafile > /opt/stack/data/ca-bundle.pem > ++ lib/keystone:configure_keystone_authtoken_middleware:461 : iniset > /etc/ceilometer/ceilometer.conf keystone_authtoken memcached_servers > localhost:11211 > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:270 > : [[ libvirt = \v\s\p\h\e\r\e ]] > ++ /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:277 > : _ceilometer_configure_storage_backend > ++ > /opt/stack/ceilometer/devstack/plugin.sh:_ceilometer_configure_storage_backend:208 > : '[' gnocchi = none ']' > ++ > /opt/stack/ceilometer/devstack/plugin.sh:_ceilometer_configure_storage_backend:210 > : '[' gnocchi = gnocchi ']' > ++ > /opt/stack/ceilometer/devstack/plugin.sh:_ceilometer_configure_storage_backend:211 > : sed -i > 's/gnocchi:\/\//gnocchi:\/\/?archive_policy=ceilometer-low\&filter_project=gnocchi_swift/' > /etc/ceilometer/event_pipeline.yaml /etc/ceilometer/pipeline.yaml > ++ > /opt/stack/ceilometer/devstack/plugin.sh:_ceilometer_configure_storage_backend:212 > : [[ ,watcher,watcher-dashboard,ceilometer,gnocchi =~ gnocchi ]] > + /opt/stack/ceilometer/devstack/plugin.sh:configure_ceilometer:1 : > exit_trap What am I missing? Any configuration parameter? Thanks in advance Christian

2 1

[TripleO] Move tripleo repos stable/ussuri to End Of Life OK?
by Marios Andreou 05 Apr '22

05 Apr '22

Hello TripleO o/ The tripleo-ci team proposes that we move the stable/ussuri branch for all tripleo repos [1] to End of Life [2]. The branch was moved to extended maintenance with [3] so we can already no longer make any new releases for ussuri/tripleo repos. Are there any objections or concerns about this? If so please speak up here or directly on the patch. I think that any patches posted to ussuri nowadays are mainly about cherrypicking back to train, rather than merging something to ussuri specifically. I have posted the proposal to move to EOL at [4] and have WF-1 to wait for this discussion. If there are no objections then once we move to EOL the tripleo-ci team will remove all Ussuri related CI, check/gate/promotions. regards, marios [1] https://releases.openstack.org/teams/tripleo.html#ussuri [2] https://docs.openstack.org/project-team-guide/stable-branches.html#maintena… [3] https://review.opendev.org/c/openstack/releases/+/817623 [4] https://review.opendev.org/c/openstack/releases/+/834049

2 5

Kubernetes cluster fails with timed out error
by Karera Tony 05 Apr '22

05 Apr '22

Hello Team, I deployed Openstack Wallaby using kolla-ansible and enabled Magnum. The Installation was Ok but when I try to deploy a k8s cluster with fedora-coreos-32.20200601.3.0 image, it fails with a timeout error. When I check the /var/log/heat-config/heat-config-script logs on the master node, I get the error below. ++ kubectl get --raw=/healthz The connection to the server localhost:8080 was refused - did you specify the right host or port? + '[' ok = '' ']' + sleep 5 ++ kubectl get --raw=/healthz The connection to the server localhost:8080 was refused - did you specify the right host or port? + '[' ok = '' ']' + sleep 5 ++ kubectl get --raw=/healthz Any idea on what could be the issue. Regards Regards Tony Karera

2 4

[cinder] Zed PTG schedule
by Rajat Dhasmana 05 Apr '22

05 Apr '22

Hello Argonauts, Zed PTG is going to start from next week i.e. 4th April, 2022 to 8th April, 2022. Kindly register if you haven't already[1]. The registration is free and acts as a count of the people participating. The Cinder team will be meeting from 5th April (Tuesday) till 8th April (Friday) in the 1300-1700 UTC time slot. I've arranged the topics as per participants availability and in the following manner[2]. The connection details are on the etherpad along with the day wise schedule of topics. I've also arranged them in a tabular structure[3] (the time allotted to each topic here might vary). Hope to see everyone next week! [1] https://openinfra-ptg.eventbrite.com/ [2] https://etherpad.opendev.org/p/zed-ptg-cinder [3] https://ethercalc.openstack.org/crz6qdm7fq0v Thanks and regards Rajat Dhasmana

1 1

[tacker] No IRC meeting on Apr 5th
by Yasufumi Ogawa 05 Apr '22

05 Apr '22

Hi, I'd like to skip IRC meeting today since it's PTG week. Cheers, Yasufumi

1 0

Constraints and docs requirements
by Michael Johnson 05 Apr '22

05 Apr '22

Hello, The Designate project recently discovered that Wallaby docs builds were failing. This started happening because jinja2 newer than is specified in the constraints file was being installed, and there was a non backward compatible change that broke the doc builds. After some debugging we have determined the reason for this is that `tox -e docs` has two steps in its installation process. The first is the dependency installation which respects constraints and then after that is the installation of the package itself (in this case Designate). This second step does not supply constraints info, and it is expected that all dependencies of the package have been preinstalled by the first step. Where we get in trouble is that the docs requirements for projects like Designate (and we suspect others) are a subset of the package requirements necessary to install the packages. This means the second step in this install process discovers that many dependencies need to be installed and this happens without constraints allowing unexpected versions to be pulled in. To resolve this issue, make sure you are including the project requirements along with the doc requirements in your tox environment dependencies section[1]. This will allow pip to install all of the required dependencies, those required for the documentation generation and those for the project, to be all installed using the upper constraints. Thanks to clarkb and fungi for helping track down this issue and composing this email. [1] https://review.opendev.org/c/openstack/designate/+/836410/1/tox.ini

2 1

Support to allow only boot from cinder volumes
by Laurent Dumont 05 Apr '22

05 Apr '22

Hey folks, I am trying to see if there are any ways to instruct Openstack to prevent the usage of local storage/ephemeral disks. There are cases where : - I don't want the added complexity of Ceph. - I don't want the added hassle of using local volumes/nfs/shared storage on the computes directly. In an ideal world, creating a VM would mean that you always have a boot-from-volume with the volume being in your chosen backend. I've seen this spec : https://blueprints.launchpad.net/nova/+spec/flavor-root-disk-none But it doesn't seem to have survived the Ocata cycle. Any thoughts? Thanks!

3 7

[CloudKitty] April 2022 - PTG Summary
by Rafael Weingärtner 05 Apr '22

05 Apr '22

Hello, CloudKitty community, Below is my summary of the CloudKitty session during the PTG meeting in April. More details registered can be found at [1]. - CloudKitty Yoga release: we started a discussion reviewing the Yoga release, where we had major interesting new features introduced, such as the reprocessing API, V2 API for the already existing V1 endpoints, and so on. Moreover, we reviewed the reviewing process of patches, and it seems to be working nicely for everybody and producing results for CloudKitty; therefore, we will maintain it. - Create scope API: there is no blueprint yet, but this is one of the requests we got. The idea is that one can create the scope (project) in CloudKitty before it is discovered by CloudKitty. Therefore, one could mark test/ignored projects for the rating before they start to be used and are picked up by CloudKitty. - ElasticSearch improvements: there are some improvements already going on, and Pierre's team is the one conducting it. - Expand the "custom_fields" in the Summary GET API to support Elastic search backends. As we discussed, this feature might be interesting to be expanded and added for other backends such as ElasticSearch. However, we still need to evaluate if that would be feasible. - Where do we go from here?: And last, but not least, we discussed the future of CloudKitty and where we want to go from our current state. The consensus is that we need to improve the default configurations to other systems such as Monasca and Prometheus to facilitate for newcomers to use it; keep users happy; continue innovating and improving CloudKitty. And, as the next step for that, we will propose a Forum panel for CloudKitty at the OpenStack summit to see if we can help people's onboarding and to have a broader discussion regarding the next big developments we will be doing in CloudKitty. Those are the main topics we covered during the PTG. If I forgot something, please let me know. Now, we just gotta work to keep evolving this awesome billing stack :) Link for the PTG Etherpad: [1] [1] https://etherpad.opendev.org/p/cloudkitty-ptg-zed -- Rafael Weingärtner

1 0

[ptl][release][stable][EM] Extended Maintenance - Victoria
by Előd Illés 05 Apr '22

05 Apr '22

Hi, As Yoga was released last week and we are in a less busy period (except the PTG :)), now it is a good time to call your attention to the following: In a month Victoria is planned to transition to Extended Maintenance phase [1] (planned date: /2022-04-27/). I have generated the list of the current *open* and *unreleased* changes in stable/victoria for every repositories [2] (where there are such patches). These lists could help the teams who are planning to do a *final* release on Victoria before moving stable/victoria branches to Extended Maintenance. Feel free to edit and extend these lists to track your progress! * next week the Release Team will tag the *latest* Victoria *releases* of repositories with *victoria-em* tag. * at the planned deadline (April 27th)//the Release Team will merge all the transition patches (even the ones without any response!) * After the transition stable/victoria will be still open for bug fixes, but there won't be official releases anymore. *NOTE*: teams, please focus on wrapping up your libraries first if there is any concern about the changes, in order to avoid broken (final!) releases! Thanks, Előd [1] https://releases.openstack.org/ [2] https://etherpad.opendev.org/p/victoria-final-release-before-em

1 0

[openstack-helm] No Meeting This Week
by Gage Hugo 04 Apr '22

04 Apr '22

Hey team, Since the PTG is this week, the weekly meeting has been cancelled. We will meet for the PTG session on April 6th 1500-1700 UTC, hope to see you there!

1 0