skypilot-nightly 1.0.0.dev20250513__py3-none-any.whl → 1.0.0.dev20250515__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- sky/__init__.py +2 -2
- sky/backends/backend.py +3 -2
- sky/backends/backend_utils.py +16 -17
- sky/backends/cloud_vm_ray_backend.py +47 -16
- sky/clouds/aws.py +11 -9
- sky/clouds/azure.py +16 -13
- sky/clouds/cloud.py +4 -3
- sky/clouds/cudo.py +3 -2
- sky/clouds/do.py +3 -2
- sky/clouds/fluidstack.py +3 -3
- sky/clouds/gcp.py +25 -9
- sky/clouds/ibm.py +12 -10
- sky/clouds/kubernetes.py +3 -2
- sky/clouds/lambda_cloud.py +6 -6
- sky/clouds/nebius.py +6 -5
- sky/clouds/oci.py +9 -7
- sky/clouds/paperspace.py +3 -2
- sky/clouds/runpod.py +9 -9
- sky/clouds/scp.py +5 -3
- sky/clouds/service_catalog/data_fetchers/fetch_gcp.py +33 -11
- sky/clouds/service_catalog/gcp_catalog.py +7 -1
- sky/clouds/vast.py +8 -7
- sky/clouds/vsphere.py +4 -2
- sky/core.py +18 -12
- sky/dashboard/out/404.html +1 -1
- sky/dashboard/out/_next/static/chunks/pages/index-6b0d9e5031b70c58.js +1 -0
- sky/dashboard/out/_next/static/{2dkponv64SfFShA8Rnw0D → jFI0Y-uJZ_XDK5IGJpKFU}/_buildManifest.js +1 -1
- sky/dashboard/out/clusters/[cluster]/[job].html +1 -1
- sky/dashboard/out/clusters/[cluster].html +1 -1
- sky/dashboard/out/clusters.html +1 -1
- sky/dashboard/out/index.html +1 -1
- sky/dashboard/out/jobs/[job].html +1 -1
- sky/dashboard/out/jobs.html +1 -1
- sky/execution.py +33 -0
- sky/jobs/recovery_strategy.py +4 -1
- sky/jobs/server/core.py +6 -12
- sky/optimizer.py +19 -13
- sky/provision/kubernetes/utils.py +26 -1
- sky/resources.py +206 -43
- sky/serve/server/core.py +0 -5
- sky/serve/spot_placer.py +3 -0
- sky/server/server.py +51 -13
- sky/skylet/log_lib.py +12 -3
- sky/skylet/log_lib.pyi +5 -0
- sky/task.py +8 -6
- sky/templates/nebius-ray.yml.j2 +3 -1
- sky/utils/cli_utils/status_utils.py +6 -5
- sky/utils/controller_utils.py +39 -43
- sky/utils/dag_utils.py +4 -2
- sky/utils/resources_utils.py +3 -0
- sky/utils/schemas.py +33 -24
- {skypilot_nightly-1.0.0.dev20250513.dist-info → skypilot_nightly-1.0.0.dev20250515.dist-info}/METADATA +1 -1
- {skypilot_nightly-1.0.0.dev20250513.dist-info → skypilot_nightly-1.0.0.dev20250515.dist-info}/RECORD +58 -58
- {skypilot_nightly-1.0.0.dev20250513.dist-info → skypilot_nightly-1.0.0.dev20250515.dist-info}/WHEEL +1 -1
- sky/dashboard/out/_next/static/chunks/pages/index-f9f039532ca8cbc4.js +0 -1
- /sky/dashboard/out/_next/static/{2dkponv64SfFShA8Rnw0D → jFI0Y-uJZ_XDK5IGJpKFU}/_ssgManifest.js +0 -0
- {skypilot_nightly-1.0.0.dev20250513.dist-info → skypilot_nightly-1.0.0.dev20250515.dist-info}/entry_points.txt +0 -0
- {skypilot_nightly-1.0.0.dev20250513.dist-info → skypilot_nightly-1.0.0.dev20250515.dist-info}/licenses/LICENSE +0 -0
- {skypilot_nightly-1.0.0.dev20250513.dist-info → skypilot_nightly-1.0.0.dev20250515.dist-info}/top_level.txt +0 -0
sky/dashboard/out/jobs.html
CHANGED
@@ -1 +1 @@
|
|
1
|
-
<!DOCTYPE html><html><head><meta charSet="utf-8"/><meta name="viewport" content="width=device-width"/><link rel="preload" href="/dashboard/skypilot.svg" as="image" fetchpriority="high"/><meta name="next-head-count" content="3"/><link rel="preload" href="/dashboard/_next/static/css/c6933bbb2ce7f4dd.css" as="style"/><link rel="stylesheet" href="/dashboard/_next/static/css/c6933bbb2ce7f4dd.css" data-n-g=""/><noscript data-n-css=""></noscript><script defer="" nomodule="" src="/dashboard/_next/static/chunks/polyfills-78c92fac7aa8fdd8.js"></script><script src="/dashboard/_next/static/chunks/webpack-830f59b8404e96b8.js" defer=""></script><script src="/dashboard/_next/static/chunks/framework-87d061ee6ed71b28.js" defer=""></script><script src="/dashboard/_next/static/chunks/main-e0e2335212e72357.js" defer=""></script><script src="/dashboard/_next/static/chunks/pages/_app-e6b013bc3f77ad60.js" defer=""></script><script src="/dashboard/_next/static/chunks/678-206dddca808e6d16.js" defer=""></script><script src="/dashboard/_next/static/chunks/312-c3c8845990db8ffc.js" defer=""></script><script src="/dashboard/_next/static/chunks/979-7bf73a4c7cea0f5c.js" defer=""></script><script src="/dashboard/_next/static/chunks/845-0ca6f2c1ba667c3b.js" defer=""></script><script src="/dashboard/_next/static/chunks/236-f49500b82ad5392d.js" defer=""></script><script src="/dashboard/_next/static/chunks/pages/jobs-a75029b67aab6a2e.js" defer=""></script><script src="/dashboard/_next/static/2dkponv64SfFShA8Rnw0D/_buildManifest.js" defer=""></script><script src="/dashboard/_next/static/2dkponv64SfFShA8Rnw0D/_ssgManifest.js" defer=""></script></head><body><div id="__next"><div class="min-h-screen bg-gray-50"><div class="fixed top-0 left-0 right-0 z-50 shadow-sm"><div class="fixed top-0 left-0 right-0 bg-white z-30 h-14 px-4 border-b border-gray-200 shadow-sm"><div class="flex items-center h-full"><div class="flex items-center space-x-4 mr-6"><a class="flex items-center px-1 pt-1 h-full" href="/dashboard"><div class="h-20 w-20 flex items-center justify-center"><img alt="SkyPilot Logo" fetchpriority="high" width="80" height="80" decoding="async" data-nimg="1" class="w-full h-full object-contain" style="color:transparent" src="/dashboard/skypilot.svg"/></div></a></div><div class="flex items-center space-x-2 md:space-x-6 mr-6"><a class="inline-flex items-center border-b-2 border-transparent hover:text-blue-600 px-1 pt-1 space-x-2" href="/dashboard/clusters"><svg class="w-4 h-4" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect width="20" height="8" x="2" y="2" rx="2" ry="2"></rect><rect width="20" height="8" x="2" y="14" rx="2" ry="2"></rect><line x1="6" x2="6.01" y1="6" y2="6"></line><line x1="6" x2="6.01" y1="18" y2="18"></line></svg><span>Clusters</span></a><a class="inline-flex items-center border-b-2 border-transparent text-blue-600 px-1 pt-1 space-x-2" href="/dashboard/jobs"><svg class="w-4 h-4" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="M16 20V4a2 2 0 0 0-2-2h-4a2 2 0 0 0-2 2v16"></path><rect width="20" height="14" x="2" y="6" rx="2"></rect></svg><span>Jobs</span></a><div class="inline-flex items-center px-1 pt-1 text-gray-400"><svg class="w-4 h-4" viewBox="0 0 423.683 423.683" width="24" height="24" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" xml:space="preserve" fill="currentColor" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><g id="SVGRepo_bgCarrier" stroke-width="0"></g><g id="SVGRepo_tracerCarrier" stroke-linecap="round" stroke-linejoin="round"></g><g id="SVGRepo_iconCarrier"><g><path d="M54.376,287.577h310.459c26.48,0,48.02-13.979,48.02-40.453c0-17.916-10.001-34.07-25.559-42.292 c-19.021-72.951-86.061-125.196-162.002-125.223v-3.431h-3.854V61.814h3.854v-9.569h-31.38v9.569h3.854v14.363h-3.854v3.431 c-75.941,0.026-142.97,52.272-161.988,125.217c-15.56,8.216-25.573,24.376-25.573,42.291 C6.36,273.597,27.896,287.577,54.376,287.577z M47.676,227.145l7.214-2.424l1.617-7.447 c13.884-64.232,71.707-110.862,137.467-110.862h31.274c65.763,0,123.582,46.63,137.473,110.862l1.607,7.447l7.223,2.424 c8.678,2.92,14.506,10.946,14.506,19.979c0,11.703-9.517,13.647-21.221,13.647H54.376c-11.7,0-21.22-1.944-21.22-13.647 C33.162,238.091,38.984,230.065,47.676,227.145z M423.683,334.602v36.836H0v-36.836h25.348v-18.418h372.99v18.418H423.683z"></path></g></g></svg><span class="ml-2">Services</span><span class="text-xs ml-2 px-1.5 py-0.5 bg-gray-100 text-gray-500 rounded">Soon</span></div></div><div class="flex items-center space-x-1 ml-auto"><a href="https://skypilot.readthedocs.io/en/latest/" target="_blank" rel="noopener noreferrer" class="inline-flex items-center px-2 py-1 text-gray-600 hover:text-blue-600 transition-colors duration-150 cursor-pointer" title="Docs"><span class="mr-1">Docs</span><svg class="w-3.5 h-3.5" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="M18 13v6a2 2 0 0 1-2 2H5a2 2 0 0 1-2-2V8a2 2 0 0 1 2-2h6"></path><polyline points="15 3 21 3 21 9"></polyline><line x1="10" y1="14" x2="21" y2="3"></line></svg></a><div class="border-l border-gray-200 h-6 mx-1"></div><a href="https://github.com/skypilot-org/skypilot" target="_blank" rel="noopener noreferrer" class="inline-flex items-center justify-center p-2 rounded-full text-gray-600 hover:bg-gray-100 transition-colors duration-150 cursor-pointer" title="GitHub"><svg class="w-5 h-5" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="currentColor"><path d="M12 0c-6.626 0-12 5.373-12 12 0 5.302 3.438 9.8 8.207 11.387.599.111.793-.261.793-.577v-2.234c-3.338.726-4.033-1.416-4.033-1.416-.546-1.387-1.333-1.756-1.333-1.756-1.089-.745.083-.729.083-.729 1.205.084 1.839 1.237 1.839 1.237 1.07 1.834 2.807 1.304 3.492.997.107-.775.418-1.305.762-1.604-2.665-.305-5.467-1.334-5.467-5.931 0-1.311.469-2.381 1.236-3.221-.124-.303-.535-1.524.117-3.176 0 0 1.008-.322 3.301 1.23.957-.266 1.983-.399 3.003-.404 1.02.005 2.047.138 3.006.404 2.291-1.552 3.297-1.23 3.297-1.23.653 1.653.242 2.874.118 3.176.77.84 1.235 1.911 1.235 3.221 0 4.609-2.807 5.624-5.479 5.921.43.372.823 1.102.823 2.222v3.293c0 .319.192.694.801.576 4.765-1.589 8.199-6.086 8.199-11.386 0-6.627-5.373-12-12-12z"></path></svg></a><a href="https://slack.skypilot.co/" target="_blank" rel="noopener noreferrer" class="inline-flex items-center justify-center p-2 rounded-full text-gray-600 hover:bg-gray-100 transition-colors duration-150 cursor-pointer" title="Slack"><svg class="w-5 h-5" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="currentColor"><path transform="scale(0.85) translate(1.8, 1.8)" d="M5.042 15.165a2.528 2.528 0 0 1-2.52 2.523A2.528 2.528 0 0 1 0 15.165a2.527 2.527 0 0 1 2.522-2.52h2.52v2.52zM6.313 15.165a2.527 2.527 0 0 1 2.521-2.52 2.527 2.527 0 0 1 2.521 2.52v6.313A2.528 2.528 0 0 1 8.834 24a2.528 2.528 0 0 1-2.521-2.522v-6.313zM8.834 5.042a2.528 2.528 0 0 1-2.521-2.52A2.528 2.528 0 0 1 8.834 0a2.528 2.528 0 0 1 2.521 2.522v2.52H8.834zM8.834 6.313a2.528 2.528 0 0 1 2.521 2.521 2.528 2.528 0 0 1-2.521 2.521H2.522A2.528 2.528 0 0 1 0 8.834a2.528 2.528 0 0 1 2.522-2.521h6.312zM18.956 8.834a2.528 2.528 0 0 1 2.522-2.521A2.528 2.528 0 0 1 24 8.834a2.528 2.528 0 0 1-2.522 2.521h-2.522V8.834zM17.688 8.834a2.528 2.528 0 0 1-2.523 2.521 2.527 2.527 0 0 1-2.52-2.521V2.522A2.527 2.527 0 0 1 15.165 0a2.528 2.528 0 0 1 2.523 2.522v6.312zM15.165 18.956a2.528 2.528 0 0 1 2.523 2.522A2.528 2.528 0 0 1 15.165 24a2.527 2.527 0 0 1-2.52-2.522v-2.522h2.52zM15.165 17.688a2.527 2.527 0 0 1-2.52-2.523 2.526 2.526 0 0 1 2.52-2.52h6.313A2.527 2.527 0 0 1 24 15.165a2.528 2.528 0 0 1-2.522 2.523h-6.313z"></path></svg></a><a href="https://github.com/skypilot-org/skypilot/issues/new" target="_blank" rel="noopener noreferrer" class="inline-flex items-center justify-center p-2 rounded-full text-gray-600 hover:bg-gray-100 transition-colors duration-150 cursor-pointer" title="Leave Feedback"><svg class="w-5 h-5" stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 24 24" xmlns="http://www.w3.org/2000/svg"><g><path fill="none" d="M0 0h24v24H0z"></path><path d="M6.455 19L2 22.5V4a1 1 0 0 1 1-1h18a1 1 0 0 1 1 1v14a1 1 0 0 1-1 1H6.455zM4 18.385L5.763 17H20V5H4v13.385zM11 13h2v2h-2v-2zm0-6h2v5h-2V7z"></path></g></svg></a></div></div></div></div><div class="transition-all duration-200 ease-in-out min-h-screen" style="padding-top:56px"><main class="p-6"><div class="flex items-center justify-between mb-4 h-5"><div class="text-base"><a class="text-sky-blue hover:underline leading-none" href="/dashboard/jobs">Managed Jobs</a></div><div class="flex items-center space-x-2"><button class="inline-flex items-center justify-center whitespace-nowrap text-sm font-medium ring-offset-background transition-colors focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring focus-visible:ring-offset-2 disabled:pointer-events-none disabled:opacity-50 hover:bg-accent h-9 rounded-md px-3 text-sky-blue hover:text-sky-blue-bright" title="Refresh"><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-rotate-cw h-4 w-4 mr-1.5"><path d="M21 12a9 9 0 1 1-9-9c2.52 0 4.93 1 6.74 2.74L21 8"></path><path d="M21 3v5h-5"></path></svg><span>Refresh</span></button></div></div><div class="relative"><div class="flex flex-col space-y-1 mb-1"><div class="flex flex-wrap items-center text-sm mb-1"><span class="mr-2 text-sm font-medium">Statuses:</span><div class="flex flex-wrap gap-2 items-center"></div></div></div><div class="rounded-lg border bg-card text-card-foreground shadow-sm"><div class="relative w-full overflow-auto"><table class="w-full caption-bottom text-base"><thead class="[&_tr]:border-b"><tr class="border-b transition-colors hover:bg-muted/50 data-[state=selected]:bg-muted"><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">ID</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Name</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">User</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Submitted</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Duration</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Status</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Resources</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Cluster</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Region</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Recoveries</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0">Details</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0">Logs</th></tr></thead><tbody class="[&_tr:last-child]:border-0"><tr class="border-b transition-colors hover:bg-muted/50 data-[state=selected]:bg-muted"><td class="p-4 align-middle [&:has([role=checkbox])]:pr-0 text-center py-6" colSpan="12"><div class="flex flex-col items-center space-y-4"><p class="text-gray-500">No active jobs</p></div></td></tr></tbody></table></div></div></div></main></div></div></div><script id="__NEXT_DATA__" type="application/json">{"props":{"pageProps":{}},"page":"/jobs","query":{},"buildId":"2dkponv64SfFShA8Rnw0D","assetPrefix":"/dashboard","nextExport":true,"autoExport":true,"isFallback":false,"scriptLoader":[]}</script></body></html>
|
1
|
+
<!DOCTYPE html><html><head><meta charSet="utf-8"/><meta name="viewport" content="width=device-width"/><link rel="preload" href="/dashboard/skypilot.svg" as="image" fetchpriority="high"/><meta name="next-head-count" content="3"/><link rel="preload" href="/dashboard/_next/static/css/c6933bbb2ce7f4dd.css" as="style"/><link rel="stylesheet" href="/dashboard/_next/static/css/c6933bbb2ce7f4dd.css" data-n-g=""/><noscript data-n-css=""></noscript><script defer="" nomodule="" src="/dashboard/_next/static/chunks/polyfills-78c92fac7aa8fdd8.js"></script><script src="/dashboard/_next/static/chunks/webpack-830f59b8404e96b8.js" defer=""></script><script src="/dashboard/_next/static/chunks/framework-87d061ee6ed71b28.js" defer=""></script><script src="/dashboard/_next/static/chunks/main-e0e2335212e72357.js" defer=""></script><script src="/dashboard/_next/static/chunks/pages/_app-e6b013bc3f77ad60.js" defer=""></script><script src="/dashboard/_next/static/chunks/678-206dddca808e6d16.js" defer=""></script><script src="/dashboard/_next/static/chunks/312-c3c8845990db8ffc.js" defer=""></script><script src="/dashboard/_next/static/chunks/979-7bf73a4c7cea0f5c.js" defer=""></script><script src="/dashboard/_next/static/chunks/845-0ca6f2c1ba667c3b.js" defer=""></script><script src="/dashboard/_next/static/chunks/236-f49500b82ad5392d.js" defer=""></script><script src="/dashboard/_next/static/chunks/pages/jobs-a75029b67aab6a2e.js" defer=""></script><script src="/dashboard/_next/static/jFI0Y-uJZ_XDK5IGJpKFU/_buildManifest.js" defer=""></script><script src="/dashboard/_next/static/jFI0Y-uJZ_XDK5IGJpKFU/_ssgManifest.js" defer=""></script></head><body><div id="__next"><div class="min-h-screen bg-gray-50"><div class="fixed top-0 left-0 right-0 z-50 shadow-sm"><div class="fixed top-0 left-0 right-0 bg-white z-30 h-14 px-4 border-b border-gray-200 shadow-sm"><div class="flex items-center h-full"><div class="flex items-center space-x-4 mr-6"><a class="flex items-center px-1 pt-1 h-full" href="/dashboard"><div class="h-20 w-20 flex items-center justify-center"><img alt="SkyPilot Logo" fetchpriority="high" width="80" height="80" decoding="async" data-nimg="1" class="w-full h-full object-contain" style="color:transparent" src="/dashboard/skypilot.svg"/></div></a></div><div class="flex items-center space-x-2 md:space-x-6 mr-6"><a class="inline-flex items-center border-b-2 border-transparent hover:text-blue-600 px-1 pt-1 space-x-2" href="/dashboard/clusters"><svg class="w-4 h-4" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect width="20" height="8" x="2" y="2" rx="2" ry="2"></rect><rect width="20" height="8" x="2" y="14" rx="2" ry="2"></rect><line x1="6" x2="6.01" y1="6" y2="6"></line><line x1="6" x2="6.01" y1="18" y2="18"></line></svg><span>Clusters</span></a><a class="inline-flex items-center border-b-2 border-transparent text-blue-600 px-1 pt-1 space-x-2" href="/dashboard/jobs"><svg class="w-4 h-4" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="M16 20V4a2 2 0 0 0-2-2h-4a2 2 0 0 0-2 2v16"></path><rect width="20" height="14" x="2" y="6" rx="2"></rect></svg><span>Jobs</span></a><div class="inline-flex items-center px-1 pt-1 text-gray-400"><svg class="w-4 h-4" viewBox="0 0 423.683 423.683" width="24" height="24" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" xml:space="preserve" fill="currentColor" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><g id="SVGRepo_bgCarrier" stroke-width="0"></g><g id="SVGRepo_tracerCarrier" stroke-linecap="round" stroke-linejoin="round"></g><g id="SVGRepo_iconCarrier"><g><path d="M54.376,287.577h310.459c26.48,0,48.02-13.979,48.02-40.453c0-17.916-10.001-34.07-25.559-42.292 c-19.021-72.951-86.061-125.196-162.002-125.223v-3.431h-3.854V61.814h3.854v-9.569h-31.38v9.569h3.854v14.363h-3.854v3.431 c-75.941,0.026-142.97,52.272-161.988,125.217c-15.56,8.216-25.573,24.376-25.573,42.291 C6.36,273.597,27.896,287.577,54.376,287.577z M47.676,227.145l7.214-2.424l1.617-7.447 c13.884-64.232,71.707-110.862,137.467-110.862h31.274c65.763,0,123.582,46.63,137.473,110.862l1.607,7.447l7.223,2.424 c8.678,2.92,14.506,10.946,14.506,19.979c0,11.703-9.517,13.647-21.221,13.647H54.376c-11.7,0-21.22-1.944-21.22-13.647 C33.162,238.091,38.984,230.065,47.676,227.145z M423.683,334.602v36.836H0v-36.836h25.348v-18.418h372.99v18.418H423.683z"></path></g></g></svg><span class="ml-2">Services</span><span class="text-xs ml-2 px-1.5 py-0.5 bg-gray-100 text-gray-500 rounded">Soon</span></div></div><div class="flex items-center space-x-1 ml-auto"><a href="https://skypilot.readthedocs.io/en/latest/" target="_blank" rel="noopener noreferrer" class="inline-flex items-center px-2 py-1 text-gray-600 hover:text-blue-600 transition-colors duration-150 cursor-pointer" title="Docs"><span class="mr-1">Docs</span><svg class="w-3.5 h-3.5" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><path d="M18 13v6a2 2 0 0 1-2 2H5a2 2 0 0 1-2-2V8a2 2 0 0 1 2-2h6"></path><polyline points="15 3 21 3 21 9"></polyline><line x1="10" y1="14" x2="21" y2="3"></line></svg></a><div class="border-l border-gray-200 h-6 mx-1"></div><a href="https://github.com/skypilot-org/skypilot" target="_blank" rel="noopener noreferrer" class="inline-flex items-center justify-center p-2 rounded-full text-gray-600 hover:bg-gray-100 transition-colors duration-150 cursor-pointer" title="GitHub"><svg class="w-5 h-5" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="currentColor"><path d="M12 0c-6.626 0-12 5.373-12 12 0 5.302 3.438 9.8 8.207 11.387.599.111.793-.261.793-.577v-2.234c-3.338.726-4.033-1.416-4.033-1.416-.546-1.387-1.333-1.756-1.333-1.756-1.089-.745.083-.729.083-.729 1.205.084 1.839 1.237 1.839 1.237 1.07 1.834 2.807 1.304 3.492.997.107-.775.418-1.305.762-1.604-2.665-.305-5.467-1.334-5.467-5.931 0-1.311.469-2.381 1.236-3.221-.124-.303-.535-1.524.117-3.176 0 0 1.008-.322 3.301 1.23.957-.266 1.983-.399 3.003-.404 1.02.005 2.047.138 3.006.404 2.291-1.552 3.297-1.23 3.297-1.23.653 1.653.242 2.874.118 3.176.77.84 1.235 1.911 1.235 3.221 0 4.609-2.807 5.624-5.479 5.921.43.372.823 1.102.823 2.222v3.293c0 .319.192.694.801.576 4.765-1.589 8.199-6.086 8.199-11.386 0-6.627-5.373-12-12-12z"></path></svg></a><a href="https://slack.skypilot.co/" target="_blank" rel="noopener noreferrer" class="inline-flex items-center justify-center p-2 rounded-full text-gray-600 hover:bg-gray-100 transition-colors duration-150 cursor-pointer" title="Slack"><svg class="w-5 h-5" xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="currentColor"><path transform="scale(0.85) translate(1.8, 1.8)" d="M5.042 15.165a2.528 2.528 0 0 1-2.52 2.523A2.528 2.528 0 0 1 0 15.165a2.527 2.527 0 0 1 2.522-2.52h2.52v2.52zM6.313 15.165a2.527 2.527 0 0 1 2.521-2.52 2.527 2.527 0 0 1 2.521 2.52v6.313A2.528 2.528 0 0 1 8.834 24a2.528 2.528 0 0 1-2.521-2.522v-6.313zM8.834 5.042a2.528 2.528 0 0 1-2.521-2.52A2.528 2.528 0 0 1 8.834 0a2.528 2.528 0 0 1 2.521 2.522v2.52H8.834zM8.834 6.313a2.528 2.528 0 0 1 2.521 2.521 2.528 2.528 0 0 1-2.521 2.521H2.522A2.528 2.528 0 0 1 0 8.834a2.528 2.528 0 0 1 2.522-2.521h6.312zM18.956 8.834a2.528 2.528 0 0 1 2.522-2.521A2.528 2.528 0 0 1 24 8.834a2.528 2.528 0 0 1-2.522 2.521h-2.522V8.834zM17.688 8.834a2.528 2.528 0 0 1-2.523 2.521 2.527 2.527 0 0 1-2.52-2.521V2.522A2.527 2.527 0 0 1 15.165 0a2.528 2.528 0 0 1 2.523 2.522v6.312zM15.165 18.956a2.528 2.528 0 0 1 2.523 2.522A2.528 2.528 0 0 1 15.165 24a2.527 2.527 0 0 1-2.52-2.522v-2.522h2.52zM15.165 17.688a2.527 2.527 0 0 1-2.52-2.523 2.526 2.526 0 0 1 2.52-2.52h6.313A2.527 2.527 0 0 1 24 15.165a2.528 2.528 0 0 1-2.522 2.523h-6.313z"></path></svg></a><a href="https://github.com/skypilot-org/skypilot/issues/new" target="_blank" rel="noopener noreferrer" class="inline-flex items-center justify-center p-2 rounded-full text-gray-600 hover:bg-gray-100 transition-colors duration-150 cursor-pointer" title="Leave Feedback"><svg class="w-5 h-5" stroke="currentColor" fill="currentColor" stroke-width="0" viewBox="0 0 24 24" xmlns="http://www.w3.org/2000/svg"><g><path fill="none" d="M0 0h24v24H0z"></path><path d="M6.455 19L2 22.5V4a1 1 0 0 1 1-1h18a1 1 0 0 1 1 1v14a1 1 0 0 1-1 1H6.455zM4 18.385L5.763 17H20V5H4v13.385zM11 13h2v2h-2v-2zm0-6h2v5h-2V7z"></path></g></svg></a></div></div></div></div><div class="transition-all duration-200 ease-in-out min-h-screen" style="padding-top:56px"><main class="p-6"><div class="flex items-center justify-between mb-4 h-5"><div class="text-base"><a class="text-sky-blue hover:underline leading-none" href="/dashboard/jobs">Managed Jobs</a></div><div class="flex items-center space-x-2"><button class="inline-flex items-center justify-center whitespace-nowrap text-sm font-medium ring-offset-background transition-colors focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-ring focus-visible:ring-offset-2 disabled:pointer-events-none disabled:opacity-50 hover:bg-accent h-9 rounded-md px-3 text-sky-blue hover:text-sky-blue-bright" title="Refresh"><svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round" class="lucide lucide-rotate-cw h-4 w-4 mr-1.5"><path d="M21 12a9 9 0 1 1-9-9c2.52 0 4.93 1 6.74 2.74L21 8"></path><path d="M21 3v5h-5"></path></svg><span>Refresh</span></button></div></div><div class="relative"><div class="flex flex-col space-y-1 mb-1"><div class="flex flex-wrap items-center text-sm mb-1"><span class="mr-2 text-sm font-medium">Statuses:</span><div class="flex flex-wrap gap-2 items-center"></div></div></div><div class="rounded-lg border bg-card text-card-foreground shadow-sm"><div class="relative w-full overflow-auto"><table class="w-full caption-bottom text-base"><thead class="[&_tr]:border-b"><tr class="border-b transition-colors hover:bg-muted/50 data-[state=selected]:bg-muted"><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">ID</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Name</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">User</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Submitted</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Duration</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Status</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Resources</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Cluster</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Region</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0 sortable whitespace-nowrap">Recoveries</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0">Details</th><th class="h-12 px-4 text-left align-middle font-medium text-[hsl(var(--text-strong))] [&:has([role=checkbox])]:pr-0">Logs</th></tr></thead><tbody class="[&_tr:last-child]:border-0"><tr class="border-b transition-colors hover:bg-muted/50 data-[state=selected]:bg-muted"><td class="p-4 align-middle [&:has([role=checkbox])]:pr-0 text-center py-6" colSpan="12"><div class="flex flex-col items-center space-y-4"><p class="text-gray-500">No active jobs</p></div></td></tr></tbody></table></div></div></div></main></div></div></div><script id="__NEXT_DATA__" type="application/json">{"props":{"pageProps":{}},"page":"/jobs","query":{},"buildId":"jFI0Y-uJZ_XDK5IGJpKFU","assetPrefix":"/dashboard","nextExport":true,"autoExport":true,"isFallback":false,"scriptLoader":[]}</script></body></html>
|
sky/execution.py
CHANGED
@@ -226,11 +226,44 @@ def _execute(
|
|
226
226
|
requested_features |= task.get_required_cloud_features()
|
227
227
|
|
228
228
|
backend = backend if backend is not None else backends.CloudVmRayBackend()
|
229
|
+
# Figure out autostop config.
|
230
|
+
# Note: Ideally this can happen after provisioning, so we can check the
|
231
|
+
# autostop config from the launched resources. Before provisioning,
|
232
|
+
# we aren't sure which resources will be launched, and different
|
233
|
+
# resources may have different autostop configs.
|
229
234
|
if isinstance(backend, backends.CloudVmRayBackend):
|
230
235
|
if down and idle_minutes_to_autostop is None:
|
231
236
|
# Use auto{stop,down} to terminate the cluster after the task is
|
232
237
|
# done.
|
233
238
|
idle_minutes_to_autostop = 0
|
239
|
+
elif not down and idle_minutes_to_autostop is None:
|
240
|
+
# No autostop config specified on command line, use the
|
241
|
+
# config from resources.
|
242
|
+
# TODO(cooperc): This should be done after provisioning, in order to
|
243
|
+
# support different autostop configs for different resources.
|
244
|
+
# Blockers:
|
245
|
+
# - Need autostop config to set requested_features before
|
246
|
+
# provisioning.
|
247
|
+
# - Need to send info message about idle_minutes_to_autostop==0 here
|
248
|
+
# - Need to check if autostop is supported by the backend.
|
249
|
+
resources = list(task.resources)
|
250
|
+
for resource in resources:
|
251
|
+
if resource.autostop_config != resources[0].autostop_config:
|
252
|
+
raise ValueError(
|
253
|
+
'All resources must have the same autostop config.')
|
254
|
+
resource_autostop_config = resources[0].autostop_config
|
255
|
+
|
256
|
+
if resource_autostop_config is not None:
|
257
|
+
if resource_autostop_config.enabled:
|
258
|
+
idle_minutes_to_autostop = (
|
259
|
+
resource_autostop_config.idle_minutes)
|
260
|
+
down = resource_autostop_config.down
|
261
|
+
else:
|
262
|
+
# Autostop is explicitly disabled, so cancel it if it's
|
263
|
+
# already set.
|
264
|
+
assert not resource_autostop_config.enabled
|
265
|
+
idle_minutes_to_autostop = -1
|
266
|
+
down = False
|
234
267
|
if idle_minutes_to_autostop is not None:
|
235
268
|
if idle_minutes_to_autostop == 0:
|
236
269
|
# idle_minutes_to_autostop=0 can cause the following problem:
|
sky/jobs/recovery_strategy.py
CHANGED
@@ -86,8 +86,11 @@ class StrategyExecutor:
|
|
86
86
|
# original task.resources
|
87
87
|
task.set_resources(type(task.resources)(new_resources_list))
|
88
88
|
if isinstance(job_recovery, dict):
|
89
|
-
|
89
|
+
name = job_recovery.pop(
|
90
90
|
'strategy', registry.JOBS_RECOVERY_STRATEGY_REGISTRY.default)
|
91
|
+
assert name is None or isinstance(name, str), (
|
92
|
+
name, 'The job recovery strategy name must be a string or None')
|
93
|
+
job_recovery_name: Optional[str] = name
|
91
94
|
max_restarts_on_errors = job_recovery.pop('max_restarts_on_errors',
|
92
95
|
0)
|
93
96
|
else:
|
sky/jobs/server/core.py
CHANGED
@@ -172,9 +172,6 @@ def launch(
|
|
172
172
|
controller_resources = controller_utils.get_controller_resources(
|
173
173
|
controller=controller,
|
174
174
|
task_resources=sum([list(t.resources) for t in dag.tasks], []))
|
175
|
-
controller_idle_minutes_to_autostop, controller_down = (
|
176
|
-
controller_utils.get_controller_autostop_config(
|
177
|
-
controller=controller))
|
178
175
|
|
179
176
|
vars_to_fill = {
|
180
177
|
'remote_user_yaml_path': remote_user_yaml_path,
|
@@ -216,15 +213,12 @@ def launch(
|
|
216
213
|
# Launch with the api server's user hash, so that sky status does not
|
217
214
|
# show the owner of the controller as whatever user launched it first.
|
218
215
|
with common.with_server_user_hash():
|
219
|
-
return execution.launch(
|
220
|
-
|
221
|
-
|
222
|
-
|
223
|
-
|
224
|
-
|
225
|
-
retry_until_up=True,
|
226
|
-
fast=True,
|
227
|
-
_disable_controller_check=True)
|
216
|
+
return execution.launch(task=controller_task,
|
217
|
+
cluster_name=controller_name,
|
218
|
+
stream_logs=stream_logs,
|
219
|
+
retry_until_up=True,
|
220
|
+
fast=True,
|
221
|
+
_disable_controller_check=True)
|
228
222
|
|
229
223
|
|
230
224
|
def queue_from_kubernetes_pod(
|
sky/optimizer.py
CHANGED
@@ -671,7 +671,7 @@ class Optimizer:
|
|
671
671
|
plan: Dict[task_lib.Task, resources_lib.Resources],
|
672
672
|
) -> float:
|
673
673
|
"""Estimates the total cost of running the DAG by the plan."""
|
674
|
-
total_cost = 0
|
674
|
+
total_cost = 0.
|
675
675
|
for node in topo_order:
|
676
676
|
resources = plan[node]
|
677
677
|
if node.time_estimator_func is None:
|
@@ -777,10 +777,13 @@ class Optimizer:
|
|
777
777
|
accelerators = resources.get_accelerators_str()
|
778
778
|
spot = resources.get_spot_str()
|
779
779
|
cloud = resources.cloud
|
780
|
-
|
780
|
+
assert cloud is not None, 'Cloud must be specified'
|
781
|
+
assert (resources.instance_type is not None), \
|
782
|
+
'Instance type must be specified'
|
783
|
+
vcpus_, mem_ = cloud.get_vcpus_mem_from_instance_type(
|
781
784
|
resources.instance_type)
|
782
785
|
|
783
|
-
def format_number(x):
|
786
|
+
def format_number(x: Optional[float]) -> str:
|
784
787
|
if x is None:
|
785
788
|
return '-'
|
786
789
|
elif x.is_integer():
|
@@ -788,8 +791,8 @@ class Optimizer:
|
|
788
791
|
else:
|
789
792
|
return f'{x:.1f}'
|
790
793
|
|
791
|
-
vcpus = format_number(
|
792
|
-
mem = format_number(
|
794
|
+
vcpus = format_number(vcpus_)
|
795
|
+
mem = format_number(mem_)
|
793
796
|
|
794
797
|
if resources.zone is None:
|
795
798
|
region_or_zone = resources.region
|
@@ -814,11 +817,12 @@ class Optimizer:
|
|
814
817
|
|
815
818
|
accelerators = resources.get_accelerators_str()
|
816
819
|
spot = resources.get_spot_str()
|
820
|
+
resources = resources.assert_launchable()
|
817
821
|
cloud = resources.cloud
|
818
|
-
|
822
|
+
vcpus_, mem_ = cloud.get_vcpus_mem_from_instance_type(
|
819
823
|
resources.instance_type)
|
820
824
|
|
821
|
-
def format_number(x):
|
825
|
+
def format_number(x: Optional[float]) -> str:
|
822
826
|
if x is None:
|
823
827
|
return '-'
|
824
828
|
elif x.is_integer():
|
@@ -826,8 +830,8 @@ class Optimizer:
|
|
826
830
|
else:
|
827
831
|
return f'{x:.1f}'
|
828
832
|
|
829
|
-
vcpus = format_number(
|
830
|
-
mem = format_number(
|
833
|
+
vcpus = format_number(vcpus_)
|
834
|
+
mem = format_number(mem_)
|
831
835
|
|
832
836
|
if resources.zone is None:
|
833
837
|
region_or_zone = resources.region
|
@@ -1195,10 +1199,12 @@ def _check_specified_clouds(dag: 'dag_lib.Dag') -> None:
|
|
1195
1199
|
all_clouds_specified.add(cloud_str)
|
1196
1200
|
|
1197
1201
|
# Explicitly check again to update the enabled cloud list.
|
1198
|
-
|
1199
|
-
|
1200
|
-
|
1201
|
-
|
1202
|
+
clouds_to_check_again = list(clouds_need_recheck -
|
1203
|
+
global_disabled_clouds)
|
1204
|
+
if len(clouds_to_check_again) > 0:
|
1205
|
+
sky_check.check_capability(sky_cloud.CloudCapability.COMPUTE,
|
1206
|
+
quiet=True,
|
1207
|
+
clouds=clouds_to_check_again)
|
1202
1208
|
enabled_clouds = sky_check.get_cached_enabled_clouds_or_refresh(
|
1203
1209
|
capability=sky_cloud.CloudCapability.COMPUTE,
|
1204
1210
|
raise_if_no_cloud_access=True)
|
@@ -1736,9 +1736,16 @@ class KubernetesInstanceType:
|
|
1736
1736
|
@staticmethod
|
1737
1737
|
def is_valid_instance_type(name: str) -> bool:
|
1738
1738
|
"""Returns whether the given name is a valid instance type."""
|
1739
|
+
# Before https://github.com/skypilot-org/skypilot/pull/4756,
|
1740
|
+
# the accelerators are appended with format "--{a}{type}",
|
1741
|
+
# e.g. "4CPU--16GB--1V100".
|
1742
|
+
# Check both patterns to keep backward compatibility.
|
1743
|
+
# TODO(romilb): Backward compatibility, remove after 0.11.0.
|
1744
|
+
prev_pattern = re.compile(
|
1745
|
+
r'^(\d+(\.\d+)?CPU--\d+(\.\d+)?GB)(--\d+\S+)?$')
|
1739
1746
|
pattern = re.compile(
|
1740
1747
|
r'^(\d+(\.\d+)?CPU--\d+(\.\d+)?GB)(--[\w\d-]+:\d+)?$')
|
1741
|
-
return bool(pattern.match(name))
|
1748
|
+
return bool(pattern.match(name)) or bool(prev_pattern.match(name))
|
1742
1749
|
|
1743
1750
|
@classmethod
|
1744
1751
|
def _parse_instance_type(
|
@@ -1755,6 +1762,11 @@ class KubernetesInstanceType:
|
|
1755
1762
|
r'^(?P<cpus>\d+(\.\d+)?)CPU--(?P<memory>\d+(\.\d+)?)GB(?:--(?P<accelerator_type>[\w\d-]+):(?P<accelerator_count>\d+))?$' # pylint: disable=line-too-long
|
1756
1763
|
)
|
1757
1764
|
match = pattern.match(name)
|
1765
|
+
# TODO(romilb): Backward compatibility, remove after 0.11.0.
|
1766
|
+
prev_pattern = re.compile(
|
1767
|
+
r'^(?P<cpus>\d+(\.\d+)?)CPU--(?P<memory>\d+(\.\d+)?)GB(?:--(?P<accelerator_count>\d+)(?P<accelerator_type>\S+))?$' # pylint: disable=line-too-long
|
1768
|
+
)
|
1769
|
+
prev_match = prev_pattern.match(name)
|
1758
1770
|
if match:
|
1759
1771
|
cpus = float(match.group('cpus'))
|
1760
1772
|
memory = float(match.group('memory'))
|
@@ -1769,6 +1781,19 @@ class KubernetesInstanceType:
|
|
1769
1781
|
accelerator_count = None
|
1770
1782
|
accelerator_type = None
|
1771
1783
|
return cpus, memory, accelerator_count, accelerator_type
|
1784
|
+
# TODO(romilb): Backward compatibility, remove after 0.11.0.
|
1785
|
+
elif prev_match:
|
1786
|
+
cpus = float(prev_match.group('cpus'))
|
1787
|
+
memory = float(prev_match.group('memory'))
|
1788
|
+
accelerator_count = prev_match.group('accelerator_count')
|
1789
|
+
accelerator_type = prev_match.group('accelerator_type')
|
1790
|
+
if accelerator_count:
|
1791
|
+
accelerator_count = int(accelerator_count)
|
1792
|
+
accelerator_type = str(accelerator_type).replace('_', ' ')
|
1793
|
+
else:
|
1794
|
+
accelerator_count = None
|
1795
|
+
accelerator_type = None
|
1796
|
+
return cpus, memory, accelerator_count, accelerator_type
|
1772
1797
|
else:
|
1773
1798
|
raise ValueError(f'Invalid instance name: {name}')
|
1774
1799
|
|