Commit Graph

759 Commits

Author SHA1 Message Date
allegroai
667c2ced3d Fix very old pip version support (<20) 2023-03-28 14:34:19 +03:00
allegroai
7f5b3c8df4 Fix None config file in session causes k8s agent to raise exception 2023-03-28 14:33:55 +03:00
allegroai
46ded2864d Fix restart feature should be tested against agent session 2023-03-28 14:33:33 +03:00
allegroai
40456be948 Black formatting
Refactor path support
2023-03-05 18:05:00 +02:00
allegroai
8d51aed679 Protect against cache folders without permission 2023-03-05 18:05:00 +02:00
allegroai
bfc4ba38cd Fix torch inside nvidia containers to use preinstalled version (i.e. ==x.y.z.* matching) 2023-03-05 18:05:00 +02:00
Niels ten Boom
3cedc104df
Add poetry cwd support (#142)
Closes #138
2023-03-05 14:19:57 +02:00
Marijan Smetko
b367c80477
Switch entrypoint shell from sh to bash (#141) 2023-02-28 21:55:16 +02:00
allegroai
262b6d3a00 Update services agent entrypoint 2023-02-05 10:40:02 +02:00
allegroai
95e996bfda Reintroduce CLEARML_AGENT_SERVICES_DOCKER_RESTART accidentally reverted by a previous merge 2023-02-05 10:34:38 +02:00
allegroai
b6d132b226 Fix build fails when target is relative path 2023-02-05 10:33:32 +02:00
allegroai
4f17a2c17d Fix K8s glue does not delete pending pods if the tasks they represent were aborted 2023-02-05 10:32:16 +02:00
allegroai
00e8e9eb5a Do not allow request exceptions (only on the initial login call) 2023-02-05 10:30:45 +02:00
allegroai
af6a77918f Fix _ is allowed in k8s label names 2023-02-05 10:29:48 +02:00
allegroai
855622fd30 Support custom service on Worker.get() calls 2023-02-05 10:29:09 +02:00
allegroai
8cd12810f3 Fix login uses GET with payload which breaks when trying to connect a server running in GCP 2023-02-05 10:28:41 +02:00
achaiah
ebb955187d
Fix agent update version (#132)
* Fix agent update version

Pip install command is missing the '=='  to execute successfully

* allow for fuzzy and direct version spec

adding logic to allow for flexible version specification

* Added regex to parse 1.2.3rc4 patterns
2023-01-08 19:10:26 +02:00
pollfly
85e1fadf9b
Fix typos (#131) 2022-12-28 19:39:59 +02:00
allegroai
249b51a31b Version bump 2022-12-13 15:29:10 +02:00
allegroai
da19ef26c4 Fix pinging running task (and change default to once a minute) 2022-12-13 15:26:26 +02:00
allegroai
f69e16ea9d Fix clearml-agent build --docker stuck on certain containers 2022-12-13 15:24:32 +02:00
allegroai
efa1f71dac Version bump to v1.5.1 2022-12-10 22:18:21 +02:00
allegroai
692cb8cf13 Update six requirements 2022-12-10 22:18:10 +02:00
allegroai
ebdc215632 Remove " from pip commands in venv 2022-12-10 20:58:30 +02:00
allegroai
b2da639582 Add CLEARML_AGENT_FORCE_SYSTEM_SITE_PACKAGES env var (default true) to allow overriding default "system_site_packages: true" behavior when running tasks in containers (docker mode and k8s-glue) 2022-12-10 20:00:46 +02:00
allegroai
71fdb43f10 Version bump to v1.5.1rc0 2022-12-07 22:09:40 +02:00
allegroai
ca2791c65e Fix pip support allowing multiple pip version constraints (by default, one for <PY3.10 and one for >=PY3.10) 2022-12-07 22:09:25 +02:00
allegroai
dd75cedaab Upgrade requirements for attrs, jsonschema, pyparsing and pyjwt 2022-12-07 22:08:15 +02:00
allegroai
669fb1a6e5 Fix using deprecated types validator argument raises an error (deprecated even before jsonschema 3.0.0 and unsupported since 4.0.0) 2022-12-07 22:07:53 +02:00
allegroai
5d517c91b5 Add agent.disable_task_docker_override configuration option to disable docker override specified in executing tasks 2022-12-07 22:07:11 +02:00
allegroai
6be75abc86 Add default output URI selection to "clearml-agent init" 2022-12-07 22:06:10 +02:00
allegroai
4c777fa2ee Version bump to v1.5.0 2022-12-05 16:42:44 +02:00
allegroai
dc5e0033c8 Remove support for kubectl run
Allow customizing pod name prefix and limit pod label
Return deleted pods from cleanup
Some refactoring
2022-12-05 11:40:19 +02:00
allegroai
3dd5973734 Filter by phase when detecting hanging pods
More debug print-outs
Use task session when possible
Push task into k8s scheduler queue only if running from the same tenant
Make sure we pass git_user/pass to the task pod
Fix cleanup command not issued when no pods exist in a multi-queue setup
2022-12-05 11:29:59 +02:00
allegroai
53d379205f Support raise_error in get_bash_output() 2022-12-05 11:26:40 +02:00
allegroai
57cde21c48 Send task.ping for executing tasks every 120 seconds (set using the agent.task_ping_interval_sec configuration option) 2022-12-05 11:22:25 +02:00
allegroai
396abf13b6 Fix get_task_session() may cause an old copy of the APIClient to be used containing a reference to the previous session 2022-12-05 11:20:32 +02:00
allegroai
6e7fb5f331 Fix sending task logs fails when agent is not running in the same tenant 2022-12-05 11:19:14 +02:00
allegroai
1d5c118b70 Fix setting CLEARML_API_DEFAULT_REQ_METHOD raises an error 2022-12-05 11:18:12 +02:00
allegroai
18612aac4d Improve configuration examples 2022-12-05 11:17:27 +02:00
allegroai
76c533a2e8 Fix access to config object 2022-11-11 13:34:17 +02:00
Niels ten Boom
9eee213683
Add option to crash agent on exception using agent.crash_on_exception configuration setting (#123) 2022-11-06 17:15:39 +02:00
allegroai
e4861fc0fb Add missing settings in clearml.conf 2022-11-06 12:36:01 +02:00
allegroai
53ef984065 Update README 2022-11-06 11:53:16 +02:00
allegroai
26e62da1a8 version bump to 1.5.0rc0 2022-10-23 13:04:00 +03:00
allegroai
d2f3614ab0 Add support for CLEARML_AGENT_DOCKER_ARGS_HIDE_ENV environment variable (see agent.hide_docker_command_env_vars config option) 2022-10-23 13:04:00 +03:00
allegroai
c6d767bd64 Make venv caching the default behavior 2022-10-23 13:04:00 +03:00
allegroai
efb06891a8 Add support for PyTorch new extra_index_url repo support. We will find the correct index url based on the cuda version, and let pip do the rest. 2022-10-23 13:04:00 +03:00
allegroai
70771b12a9 Remove unused code 2022-10-23 13:04:00 +03:00
allegroai
3f7a4840cc Add support for operator != in package version (mostly for pytorch resolving) 2022-10-23 13:04:00 +03:00