This commit does a few things:
- Add the `--squash` option to the `awx-devel-build` make target. This reduces the resulting image size from 2.12 GB to 1.37 GB. I think we can get this down even more by inspecting the image contents.
- Reorganize commands so that the cache expires less often. Before this commit, any changes to the Makefile would essentially cause the entire image to rebuild.
- Break yum dependencies up into multiple lines. This makes it easier to see what changes in a diff.
- Use `n` to install our required version of node (rather `curl node | bash`). I’ve found this to be easier to maintain / more portable when working with other Dockerfiles.
- General organizational changes to make things easier to parse visually.
Installing the latest python36-setuptools automatically creates the symlink from python3 -> python3.6 and from python36 -> /usr/bin/python3.6. Building the images fails when the symlink is created explicitly in the AWX installer.
Signed-off-by: Uriel Mandujano <uriel.mandujano14@gmail.com>
Daphne has a default timeout of 86400 seconds, so after 1 day of starting
awx_web container, the stdout stops refreshing automatically on the web UI.
This fixes this issue by making the timeout infinite, so the connection
between nginx and daphne's websocket never closes.
run this command on _any_ node in an awx cluster:
$ awx-manage profile_sql --threshold=2.0 --minutes=1
...and for 1 minute, the timing for _every_ SQL query in _every_ awx
Python process that uses the Django ORM will be measured
queries that run longer than (in this example) 2 seconds will be
written to a per-process sqlite database in /var/lib/awx/profile, and
the file will contain an EXPLAIN VERBOSE for the query and the full
Python stack that led to that SQL query's execution (this includes not
just WSGI requests, but background processes like the runworker and
dispatcher)
$ awx-manage profile_sql --threshold=0
...can be used to disable profiling again (if you don't want to wait for
the minute to expire)
this commit implements the bulk of `awx-manage run_dispatcher`, a new
command that binds to RabbitMQ via kombu and balances messages across
a pool of workers that are similar to celeryd workers in spirit.
Specifically, this includes:
- a new decorator, `awx.main.dispatch.task`, which can be used to
decorate functions or classes so that they can be designated as
"Tasks"
- support for fanout/broadcast tasks (at this point in time, only
`conf.Setting` memcached flushes use this functionality)
- support for job reaping
- support for success/failure hooks for job runs (i.e.,
`handle_work_success` and `handle_work_error`)
- support for auto scaling worker pool that scale processes up and down
on demand
- minimal support for RPC, such as status checks and pool recycle/reload
