External-Mirrors/awx
2
0
Fork 0
mirror of https://github.com/ansible/awx.git synced 2026-01-31 17:18:59 -03:30
Code Issues Packages Projects Releases Wiki Activity

Update some contributing docs

Browse Source 
* Update the tools called in the dev environment
* More RMQ purges from architecture docs
* Remove the old clusterdev target
This commit is contained in:
Matthew Jones
2020-04-08 10:02:05 -04:00
parent d40143a63d
commit 6abeaf2c55
12 changed files with 12 additions and 527 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docs/task_manager_system.md
View File

@@ -26,7 +26,7 @@ Empirically, the periodic task manager has been effective in the past and will c
* Detect finished workflow jobs
* Spawn next workflow jobs if needed
* For each pending job, start with the oldest created job
* If the job is not blocked, and there is capacity in the instance group queue, then mark it as `waiting` and submit the job to RabbitMQ.
* If the job is not blocked, and there is capacity in the instance group queue, then mark it as `waiting` and submit the job.
### Job Lifecycle

4
docs/tasks.md
View File

@@ -26,9 +26,7 @@ Tasks, Queues and Workers
To accomplish this, AWX makes use of a "Task Queue" abstraction. Task Queues are used as a mechanism to distribute work across machines in an AWX installation. A Task Queue's input is a unit of work called a Task. Dedicated worker processes running on every AWX node constantly monitor these queues for new work to perform.
AWX communicates with these worker processes to mediate between clients and workers. This is done via distributed RabbitMQ queues and the already-acknowledged local queue that the Dispatcher is working through. Simply put: to initiate a task, the client (generally, Python code in the AWX API) publishes a message to a queue, and RabbitMQ then delivers that message to one or more workers.
By default, when AWX creates queues in RabbitMQ, it creates them as *durable* queues in RabbitMQ (which allows for message persistence at the cost of lower performance). For increased message throughput (at the risk of message loss on server restarts), set BROKER_DURABILITY=False, and AWX will create _transient_ queues instead.
AWX communicates with these worker processes to mediate between clients and workers. This is done via distributed queues and the already-acknowledged local queue that the Dispatcher is working through. Simply put: to initiate a task, the client (generally, Python code in the AWX API) publishes a message to a queue, which is then delivered to one or more workers.
Clustered AWX installations consist of multiple workers spread across every
node, giving way to high availability and horizontal scaling.