| opendevreview | OpenStack Proposal Bot proposed openstack/project-config master: Normalize projects.yaml https://review.opendev.org/c/openstack/project-config/+/940092 | 02:40 |
|---|---|---|
| opendevreview | Merged openstack/project-config master: Normalize projects.yaml https://review.opendev.org/c/openstack/project-config/+/940092 | 07:18 |
| opendevreview | Dr. Jens Harbott proposed openstack/project-config master: Revert "Use ubuntu-noble for test-release-openstack" https://review.opendev.org/c/openstack/project-config/+/940105 | 08:08 |
| *** tosky_ is now known as tosky | 08:53 | |
| opendevreview | Stephen Finucane proposed openstack/pbr master: Add openstack-tox-py313 job https://review.opendev.org/c/openstack/pbr/+/940117 | 10:30 |
| opendevreview | Merged openstack/project-config master: Revert "Use ubuntu-noble for test-release-openstack" https://review.opendev.org/c/openstack/project-config/+/940105 | 10:49 |
| *** tosky is now known as Guest6863 | 13:19 | |
| *** tosky_ is now known as tosky | 13:19 | |
| opendevreview | Merged openstack/project-config master: [Neutron] Update dashboard with latest job renames https://review.opendev.org/c/openstack/project-config/+/940065 | 13:56 |
| opendevreview | Jeremy Stanley proposed openstack/pbr master: Add openstack-tox-py313 job https://review.opendev.org/c/openstack/pbr/+/940117 | 17:07 |
| gboutry | Good day, I've been looking into accessing bigger runners to solve our CI stability issues, but I'm getting rejected by zuul about nodeset ubuntu-noble-32GB not existing. Is something specific needed at admin level to gain access? | 18:17 |
| clarkb | gboutry: no there shouldn't be restrictions on that. However, that lable is only provided by a single provider and it may not be functional | 18:19 |
| clarkb | are you getting NODE_FAILURE errors? | 18:19 |
| gboutry | The nodeset "ubuntu-noble-32GB" was not found. | 18:20 |
| clarkb | oh we probably node have a nodeset defined for it so you have to define your own | 18:20 |
| clarkb | the label is in nodepool though | 18:20 |
| clarkb | as an aside keep in mind that because those labels are only provided by one cloud you're at the mercy of its uptime and its more restricted quota limits. Using the larger labels like that should be more tactical where absolutely necessary rather than treated as a default | 18:21 |
| fungi | https://opendev.org/openstack/openstack-ansible-os_magnum/src/branch/master/zuul.d/nodesets.yaml#L7-L11 is an example of doing that | 18:21 |
| clarkb | you can also do it as part of the job itself: https://opendev.org/opendev/system-config/src/branch/master/zuul.d/system-config-run.yaml#L60-L70 | 18:22 |
| gboutry | Oh, I didn't get the difference between nodepool and nodeset, thanks for the example. | 18:23 |
| clarkb | out of curiousity what sort of stability problems are you having? Are they issues arising due to diving into swap? | 18:23 |
| fungi | technically the difference is between labels and nodesets | 18:24 |
| clarkb | labels represent that actual resource in a cloud where it may be. Then a nodeset is an aggregation of labels/resources giving them logical names for use in ansible | 18:25 |
| fungi | a label is an identifier for a specific type of node. a nodeset is a group of one or more nodes (technically zero or more since you can have an empty nodeset i think), they can be multiples of the same label or a variety of labels | 18:25 |
| gboutry | We have 4 jobs running functional tests, but during the part where there's most activity, the kubeapi server keeps dying due to resource constraints. | 18:25 |
| fungi | oom kills? | 18:25 |
| clarkb | controlling the amount of work that executes in parallel, ensuring there is at least some swap present on the node, and fixing inefficient resource utilization /me looks at privsep are all options that can be considered too | 18:27 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!