| -@gerrit:opendev.org- Dong Zhang proposed on behalf of Benedikt Löffler: [zuul/nodepool] 923506: Support EC2 Fleet API https://review.opendev.org/c/zuul/nodepool/+/923506 | 14:18 | |
| -@gerrit:opendev.org- Dong Zhang proposed on behalf of Benedikt Löffler: [zuul/nodepool] 923506: Support EC2 Fleet API https://review.opendev.org/c/zuul/nodepool/+/923506 | 14:20 | |
| -@gerrit:opendev.org- Dong Zhang proposed on behalf of Benedikt Löffler: [zuul/nodepool] 923506: Support EC2 Fleet API https://review.opendev.org/c/zuul/nodepool/+/923506 | 14:23 | |
| -@gerrit:opendev.org- Dong Zhang proposed on behalf of Benedikt Löffler: [zuul/nodepool] 923506: Support EC2 Fleet API https://review.opendev.org/c/zuul/nodepool/+/923506 | 14:27 | |
| -@gerrit:opendev.org- Dong Zhang proposed on behalf of Benedikt Löffler: [zuul/nodepool] 923506: Support EC2 Fleet API https://review.opendev.org/c/zuul/nodepool/+/923506 | 14:41 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed on behalf of Benedikt Löffler: [zuul/nodepool] 923506: Support EC2 Fleet API https://review.opendev.org/c/zuul/nodepool/+/923506 | 16:28 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed on behalf of Benedikt Löffler: [zuul/nodepool] 923506: Support EC2 Fleet API https://review.opendev.org/c/zuul/nodepool/+/923506 | 16:38 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 17:46 | |
| - [zuul/zuul] 924447: Move some attributes from label to flavor https://review.opendev.org/c/zuul/zuul/+/924447 | ||
| - [zuul/zuul] 924448: Add abstract statemachine adapter methods to provider https://review.opendev.org/c/zuul/zuul/+/924448 | ||
| - [zuul/zuul] 924449: Split awsprovider.py https://review.opendev.org/c/zuul/zuul/+/924449 | ||
| - [zuul/zuul] 924668: Log the reason we fail a nodeset request https://review.opendev.org/c/zuul/zuul/+/924668 | ||
| - [zuul/zuul] 924669: Remove nodepool part of single-tenant fixture https://review.opendev.org/c/zuul/zuul/+/924669 | ||
| - [zuul/zuul] 924670: Remove extra nesting in test_launcher https://review.opendev.org/c/zuul/zuul/+/924670 | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 18:55 | |
| - [zuul/zuul] 924669: Remove nodepool part of single-tenant fixture https://review.opendev.org/c/zuul/zuul/+/924669 | ||
| - [zuul/zuul] 924670: Remove extra nesting in test_launcher https://review.opendev.org/c/zuul/zuul/+/924670 | ||
| - [zuul/zuul] 924447: Move some attributes from label to flavor https://review.opendev.org/c/zuul/zuul/+/924447 | ||
| - [zuul/zuul] 924448: Add abstract statemachine adapter methods to provider https://review.opendev.org/c/zuul/zuul/+/924448 | ||
| - [zuul/zuul] 924449: Split awsprovider.py https://review.opendev.org/c/zuul/zuul/+/924449 | ||
| @yodakv:matrix.org | Hello everyone, | 20:20 |
|---|---|---|
| Does anybody know if node pool Metastatic driver can work with multiple launcher instances ? I've strange issue , only when launcher instance are more that one. | ||
| With single instance everything work as expected. I use the same configuration example from documentation , maybe isn't enough ? | ||
| Metadata is shared well, all launchers can see it, so does can be an configuration issues ? | ||
| Let try to describe my issue: | ||
| On first job execution attempt the zuul-launcher-0 instance create node + backing_node record in zk, after job execution the node is removed in zk, the backing_node is locked, in-use state and available for new job. | ||
| On second job execution attempt using same job from previous step, the node request is handled by the zuul-launcher-1 instance, expected behaver would reusing backing_node created from previous request , but unfortunately doesn't happens. The zuul-launcher-1 instance create new node + backing_node, after job is executed, node is removed form zk, new created backing_node is locked , in-use state. | ||
| And most important part , created backing_nodes are never delete/deallocate, does that means the launcher lose some state ? | ||
| I use Zuul 9.5 and nodepool 9.1 . Any advice ? | ||
| @yodakv:matrix.org | Is it possible to be bug ? | 20:23 |
| @clarkb:matrix.org | Kiril Valchev: I think generally the rule is to not share providers across launchers. With metastatic I suspect this applies to the underlying providers as well. | 20:24 |
| @clarkb:matrix.org | is it possible you have multiple launchers using the same underlying provider for metastatic? | 20:24 |
| @clarkb:matrix.org | corvus: I assume the jaeger thing is still a problem? https://review.opendev.org/c/zuul/zuul/+/923975 seems like something that we should probably have in quickstart, but I also suspect we either need to hold a test node or build a quickstart env locally because the logs aren't super helpful in this case | 20:30 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 924682: WIP: Drive statemachine in launcher https://review.opendev.org/c/zuul/zuul/+/924682 | 22:27 | |
| @jim:acmegating.com | Clark: i haven't done or seen anything to address the jaeger thing, so i assume it's status quo (broken on latest). that change lgtm; i think you can approve if you want or stack on it if you want to get logs from latest; running locally is not hard, btw: | 22:32 |
| ``` | ||
| cd zuul/playbooks/tutorial | ||
| ansible-playbook run-localtest-quick-start.yaml | ||
| ``` | ||
| @dfajfer:fsfe.org | hey guys, why is web.root here https://zuul-ci.org/docs/zuul/latest/configuration.html#attr-web listed as required? I've been running without that fine | 11:27 |
| The only caveat is how comments are coming on Gerrit (I just tested this on Gerrit): | ||
| `build-docker build/a8fd5d2fea934de1b1e69abe728df2c3 : SUCCESS in 4m 00s` | ||
| Instead of the usual: | ||
| `build-docker https://my.zuul/t/tenant/build/a8fd5d2fea934de1b1e69abe728df2c3 : SUCCESS in 4m 00s` | ||
| @dfajfer:fsfe.org | > <@dfajfer:fsfe.org> hey guys, why is web.root here https://zuul-ci.org/docs/zuul/latest/configuration.html#attr-web listed as required? I've been running without that fine | 11:28 |
| > | ||
| > The only caveat is how comments are coming on Gerrit (I just tested this on Gerrit): | ||
| > | ||
| > `build-docker build/a8fd5d2fea934de1b1e69abe728df2c3 : SUCCESS in 4m 00s` | ||
| > | ||
| > Instead of the usual: | ||
| > | ||
| > `build-docker https://my.zuul/t/tenant/build/a8fd5d2fea934de1b1e69abe728df2c3 : SUCCESS in 4m 00s` | ||
| Documentation suggests that `ewb.status_url` should be handling that, but `tenant.web-root` looks to be doing the same as `web.status_url` but it does replace `web.root` instead | ||
| @dfajfer:fsfe.org | * Documentation suggests that `web.status_url` should be handling that, but `tenant.web-root` looks to be doing the same as `web.status_url` but it does replace `web.root` instead | 11:28 |
| @dfajfer:fsfe.org | am I missing something as an operator user or it's all wrong? | 11:28 |
| @dfajfer:fsfe.org | CRD also ignores `web.root` entirely and just allows to provide `web.status_url` which isn't doing anything really but I'll fix that | 11:29 |
| @dfajfer:fsfe.org | * zuul-operator CRD also ignores `web.root` entirely and just allows to provide `web.status_url` which isn't doing anything really but I'll fix that | 11:29 |
| @dfajfer:fsfe.org | ```ini | 11:34 |
| [web] | ||
| listen_address=0.0.0.0 | ||
| port=9000 | ||
| ``` | ||
| This is the entire operator config for web so it's expected that setting root/status_url doesn't do anything in the operator | ||
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!