BM1: the build modules give a step met away again by again-delays
Each build module declares again-delays (15s,60s) and gives its puller the delays steprun.AgainDelays reads from it; a publish's no responders is a transport fault by steprun.PublishFault with no seat call to learn why, so the modules no longer invoke the assigner's assignments. Builds: design 47 H2 §799; H-A-350 DMC-7, H-A-390
This commit is contained in:
@@ -46,6 +46,13 @@
|
||||
"description": "How many processors one Node step of this module needs (tsc, esbuild and a test runner; a median step takes 15 s). The capacity above is an upper bound: on each node the module runs at most floor(the cores of its build slice's CPU quota / this) steps at once, the quota read live from its cgroup (issue 1806), so a lowered build CPU share never starves its steps. Above 0.",
|
||||
"minimum": 0.1,
|
||||
"default": 2
|
||||
},
|
||||
"again-delays": {
|
||||
"type": "string",
|
||||
"description": "The delays a step ask whose fetch or publish met its holder away is given again after: after its first run, then after its second, as durations such as 15s,60s (design 47 H2 §799); a third run met so ends the step failed with why fetch: or publish:.",
|
||||
"trusted": false,
|
||||
"runtime": true,
|
||||
"default": "15s,60s"
|
||||
}
|
||||
},
|
||||
"tools": [
|
||||
@@ -79,7 +86,6 @@
|
||||
"seat:oci-artifact-registry.publish-part",
|
||||
"seat:npm-package-registry.publish",
|
||||
"seat:npm-package-registry.publish-part",
|
||||
"seat:mesh-module-assigner.assignments",
|
||||
"seat:mesh-catalog.module",
|
||||
"seat:mesh-catalog.seat-definitions"
|
||||
],
|
||||
|
||||
Reference in New Issue
Block a user