Chris Meyers
c4fb88c0d9
remove uneeded post commit wrapper
...
* Since we changed the lower level method to always use post commit
message emit
2017-02-28 13:10:00 -05:00
Chris Meyers
dd513621f0
Revert "Merge pull request #5553 from chrismeyersfsu/fix-waiting_blocked"
...
This reverts commit 9ba2122f4f85eecaeb6fa53ac92ea2811b05e83f, reversing
changes made to c3a5f2c96fd85dd1405a8f5c875ffc988dee16a4.
2017-02-27 09:38:45 -05:00
Chris Meyers
903d0472f0
just like we fail running tasks fail waiting tasks
...
* Associate the celery_id with the job at the earliest point possible.
This ensures that a waiting job has a celery id. Thus, we are free to
fail waiting jobs that don't have a celery id.
2017-02-24 12:07:04 -05:00
Chris Meyers
08825a1f49
fix check running status
2017-02-23 15:09:50 -05:00
Chris Meyers
6e9488a59b
ensure job deps are created only once
2017-02-15 15:54:30 -05:00
Chris Meyers
dfb7b08380
check job status before marking as failed
...
* When rectifying celery <-> db job status, re-check the job status to
ensure it did not finish between the time that we pulled the list of
celery tasks from celery and the time in which we rectify.
2017-02-09 15:27:08 -05:00
AlanCoding
b54de150db
corner-case verbose message for users who do not provide an inventory for their job
2017-01-31 12:49:26 -05:00
AlanCoding
a8e46021f8
conservative method to validating that a job can be launched from a node
2017-01-18 12:16:45 -05:00
AlanCoding
5bfb0f229c
graceful handling of null project and inventory in workflow
2017-01-17 14:19:46 -05:00
AlanCoding
8f259f2bb7
handle case of node with null UJT
2017-01-10 18:50:52 -05:00
Chris Meyers
4509f402a8
send notification when recifiying celery jobs
2016-12-15 16:31:39 -05:00
Chris Meyers
812b7c5f5f
fix two inv updates created from 1 jt run
2016-12-09 14:09:58 -05:00
Chris Meyers
cc7c2957cf
always chain failures
...
* When inv and proj updates trigger from a JT run, if either update
fails then the job template should get marked failed. Before this
commit, the job template would get marked failed ONLY if there was
enough capacity to run all the associated updates within the same
schedule() call. If, instead, the associated updates were ran in another
schedule() call, the failure chain was lost. This changeset fixes that
by saving the necessary data in the dependent_jobs relationship so that
the failure is always chained.
2016-12-09 14:09:54 -05:00
Aaron Tan
a458bb3c89
Remove source_workflow_job and use workflow_job_id instead.
2016-11-29 11:55:04 -05:00
Aaron Tan
54f99749eb
Provide linkage from spawned job back to wfj.
2016-11-25 21:23:29 -05:00
Aaron Tan
9e4655419e
Fix flake8 E302 errors.
2016-11-15 20:59:39 -05:00
Chris Meyers
9aaf694acf
always allow running at least 1 job
...
* Regardless of capacity
* Add unit tests around this logic
2016-11-10 11:07:05 -05:00
Aaron Tan
cdbc5127a6
Workflow notification backend.
2016-11-08 15:55:15 -05:00
Chris Meyers
930b3651a0
Merge pull request #3917 from chrismeyersfsu/task-manager_prov_callback_inv_sync_dedup
...
on host callback, upate dynamic inv only once
2016-11-08 15:45:38 -05:00
Chris Meyers
adcf0af2a0
on host callback, upate dynamic inv only once
...
* During a job template launch via the host callback launch feature,
when the host launching the job template is not found; we try to launch
associated dyn inv updates to pick up the host. Then the inv is searched
again for the host. If the host is found, then the jt is launched. This
fix carries forward the feature that prevents the dyn inv update with
cache timeout of 0 from being re-launched with the new task manager.
2016-11-08 15:06:47 -05:00
Aaron Tan
05790c5b5c
Merge pull request #3848 from jangsutsr/3528_workflow_job_cancel_endpoint
...
Implement workflow job cancel endpoint
2016-11-08 12:16:39 -05:00
Aaron Tan
9e244b640b
Simplify internal node job cancel mechanism.
2016-11-08 11:32:33 -05:00
Aaron Tan
3c0d60075c
Fix bug by using cancel_flag.
2016-11-08 11:32:33 -05:00
Aaron Tan
3778914aa9
Implement recursive spawned job cancel.
2016-11-08 11:31:06 -05:00
Chris Meyers
dcb8789fd9
fix job blocking job with allow simultaneous
...
related to #3911
2016-11-08 11:00:52 -05:00
Matthew Jones
343966f744
Implement gathering overall task capacity
...
For use when running/planning jobs
2016-11-07 13:45:01 -05:00
Matthew Jones
8e77deea27
Push celery queue stats to memcached periodically
...
We'll piggyback off the task that checks for inconsistent jobs between
celery and Tower itself. These are read off via /api/v1/ping
2016-11-03 12:10:38 -04:00
Chris Meyers
87dd91e849
rename Scheduler to TaskManager
2016-11-01 13:50:42 -05:00
Chris Meyers
c4a3b604f8
add back in alan's workflow fail detection
2016-11-01 13:50:42 -05:00
Chris Meyers
ed37e68c53
run dependencies when capacity is available
2016-11-01 13:50:42 -05:00
Chris Meyers
0f98e1edec
remove todo's
2016-11-01 13:50:42 -05:00
Chris Meyers
454b3edb7c
rectify celery<->db inconsistent running job
2016-11-01 13:50:42 -05:00
Chris Meyers
4ef4b4709b
workflow execution added
2016-11-01 13:50:42 -05:00
Chris Meyers
9802b1f379
AdHocCommand support added to task manager
2016-11-01 13:50:42 -05:00
Chris Meyers
46faeffbb3
added task manager system job support
2016-11-01 13:50:42 -05:00
Chris Meyers
306562cd67
inventory updates running correctly
2016-11-01 13:50:42 -05:00
Chris Meyers
555f0bb90f
project and jobs running correctly
2016-11-01 13:50:42 -05:00
AlanCoding
a2c972e513
Workflow status original commit
2016-10-27 12:52:07 -04:00
Alan Rominger
76fae6c052
Revert "Introduce workflow failure condition"
2016-10-27 12:21:31 -04:00
AlanCoding
56a9978d56
Introduce workflow failure condition
2016-10-25 09:14:52 -04:00
Wayne Witzel III
9049fb208f
Merge branch 'devel' of https://github.com/ansible/ansible-tower into jtabor-sockets
2016-10-06 10:06:30 -04:00
Wayne Witzel III
71c84a6383
convert scheduler to use websocket
2016-10-06 09:58:51 -04:00
Wayne Witzel III
6d4a2a8f8e
converting from socketio to channels websocket
2016-10-04 14:13:16 -07:00
AlanCoding
2faafa8071
Merge branch 'devel' of https://github.com/ansible/ansible-tower into wf_rbac_prompt
2016-09-30 10:06:23 -04:00
Chris Meyers
fbc1dff4de
flake8 fixes
2016-09-29 10:23:00 -04:00
Chris Meyers
0d538f8b0f
Revert "removed wait_task look restriction"
...
This reverts commit f159fd45406c6af2aacc48a44e2a3993bfa19ce8.
2016-09-28 10:23:45 -04:00
Chris Meyers
89250dcf36
removed wait_task look restriction
2016-09-27 16:04:00 -04:00
Chris Meyers
3a8033dec4
cheesy global lock
2016-09-27 14:16:18 -04:00
Chris Meyers
04f69727f2
fully message driven job execution
...
TODO:
* Need a distributed lock (leverage postgres)
* Less memory-intensive graph representation
* Maybe serializer/deserializer graph to database
* Iterative graph building instead of full rebuild.
2016-09-27 14:16:18 -04:00
Chris Meyers
cdb65ccac9
replace task manager with event driven scheduler
2016-09-27 14:16:18 -04:00
