Deploying Nautobot: Web Service and Workers¶
Like most Django applications, Nautobot runs as a WSGI application behind an HTTP server.
Additionally, certain Nautobot features (including Git repository synchronization, Webhooks, Jobs, etc.) depend on the presence of Nautobot's Celery background worker process, which is not automatically started with Nautobot and is run as a separate service.
This document will guide you through setting up uWSGI and establishing Nautobot web and Celery worker services to run on system startup.
Nautobot includes a
nautobot-server start management command that directly invokes uWSGI. This command behaves exactly as uWSGI does, but allows us to maintain a single entrypoint into the Nautobot application.
Nautobot requires at least one worker to consume background tasks required for advanced background features. A
nautobot-server celery command is included that directly invokes Celery. This command behaves exactly as the Celery command-line utility does, but launches it through Nautobot's environment to share Redis and database connection settings transparently.
Changed in version 1.1.0
Prior to version 1.1.0, Nautobot utilized RQ as the primary background task worker. As of Nautobot 1.1.0, RQ is now deprecated. RQ and the
@job decorator for custom tasks are still supported for now, but users should migrate the primary worker to Celery and then, only if still required, run RQ concurrently with the Celery worker. RQ and the
@job decorator will no longer be documented, and support for RQ will be removed in a future release.
Advanced Task Queue Configuration¶
You may want to deploy multiple workers and/or multiple queues. For more information see the task queues documentation.
nautobot user, copy and paste the following into
[uwsgi] ; The IP address (typically localhost) and port that the WSGI process should listen on socket = 127.0.0.1:8001 ; Fail to start if any parameter in the configuration file isn’t explicitly understood by uWSGI strict = true ; Enable master process to gracefully re-spawn and pre-fork workers master = true ; Allow Python app-generated threads to run enable-threads = true ;Try to remove all of the generated file/sockets during shutdown vacuum = true ; Do not use multiple interpreters, allowing only Nautobot to run single-interpreter = true ; Shutdown when receiving SIGTERM (default is respawn) die-on-term = true ; Prevents uWSGI from starting if it is unable load Nautobot (usually due to errors) need-app = true ; By default, uWSGI has rather verbose logging that can be noisy disable-logging = true ; Assert that critical 4xx and 5xx errors are still logged log-4xx = true log-5xx = true ; Enable HTTP 1.1 keepalive support http-keepalive = 1 ; ; Advanced settings (disabled by default) ; Customize these for your environment if and only if you need them. ; Ref: https://uwsgi-docs.readthedocs.io/en/latest/Options.html ; ; Number of uWSGI workers to spawn. This should typically be 2n+1, where n is the number of CPU cores present. ; processes = 5 ; If using subdirectory hosting e.g. example.com/nautobot, you must uncomment this line. Otherwise you'll get double paths e.g. example.com/nautobot/nautobot/. ; Ref: https://uwsgi-docs.readthedocs.io/en/latest/Changelog-2.0.11.html#fixpathinfo-routing-action ; route-run = fixpathinfo: ; If hosted behind a load balancer uncomment these lines, the harakiri timeout should be greater than your load balancer timeout. ; Ref: https://uwsgi-docs.readthedocs.io/en/latest/HTTP.html?highlight=keepalive#http-keep-alive ; harakiri = 65 ; add-header = Connection: Keep-Alive ; http-keepalive = 1
This configuration should suffice for most initial installations, you may wish to edit this file to change the bound IP address and/or port number, or to make performance-related adjustments. See uWSGI documentation for the available configuration parameters.
If you are deploying uWSGI behind a load balancer be sure to configure the harakiri timeout and keep alive appropriately.
systemd to control both uWSGI and Nautobot's background worker processes.
The following steps must be performed with root permissions.
First, we'll establish the
systemd unit file for the Nautobot web service. Copy and paste the following into
[Unit] Description=Nautobot WSGI Service Documentation=https://docs.nautobot.com/projects/core/en/stable/ After=network-online.target Wants=network-online.target [Service] Type=simple Environment="NAUTOBOT_ROOT=/opt/nautobot" User=nautobot Group=nautobot PIDFile=/var/tmp/nautobot.pid WorkingDirectory=/opt/nautobot ExecStart=/opt/nautobot/bin/nautobot-server start --pidfile /var/tmp/nautobot.pid --ini /opt/nautobot/uwsgi.ini ExecStop=/opt/nautobot/bin/nautobot-server start --stop /var/tmp/nautobot.pid ExecReload=/opt/nautobot/bin/nautobot-server start --reload /var/tmp/nautobot.pid Restart=on-failure RestartSec=30 PrivateTmp=true [Install] WantedBy=multi-user.target
Nautobot Background Services¶
Changed in version 1.1.0
Prior to version 1.1.0, Nautobot utilized RQ as the primary background task worker. As of Nautobot 1.1.0, RQ is now deprecated and has been replaced with Celery. RQ can still be used by plugins for now, but will be removed in a future release. Please migrate your deployment to utilize Celery as documented below.
Next, we will setup the
systemd units for the Celery worker and Celery Beat scheduler.
Added in version 1.1.0
The Celery worker service consumes tasks from background task queues and is required for taking advantage of advanced Nautobot features including Jobs, Custom Fields, and Git Repositories, among others.
To establish the
systemd unit file for the Celery worker, copy and paste the following into
[Unit] Description=Nautobot Celery Worker Documentation=https://docs.nautobot.com/projects/core/en/stable/ After=network-online.target Wants=network-online.target [Service] Type=exec Environment="NAUTOBOT_ROOT=/opt/nautobot" User=nautobot Group=nautobot PIDFile=/var/tmp/nautobot-worker.pid WorkingDirectory=/opt/nautobot ExecStart=/opt/nautobot/bin/nautobot-server celery worker --loglevel INFO --pidfile /var/tmp/nautobot-worker.pid Restart=always RestartSec=30 PrivateTmp=true [Install] WantedBy=multi-user.target
Celery Beat Scheduler¶
Added in version 1.2.0
The Celery Beat scheduler enables the periodic execution of and scheduling of background tasks. It is required to take advantage of the job scheduling and approval features.
To establish the
systemd unit file for the Celery Beat scheduler, copy and paste the following into
[Unit] Description=Nautobot Celery Beat Scheduler Documentation=https://docs.nautobot.com/projects/core/en/stable/ After=network-online.target Wants=network-online.target [Service] Type=exec Environment="NAUTOBOT_ROOT=/opt/nautobot" User=nautobot Group=nautobot PIDFile=/var/tmp/nautobot-scheduler.pid WorkingDirectory=/opt/nautobot ExecStart=/opt/nautobot/bin/nautobot-server celery beat --loglevel INFO --pidfile /var/tmp/nautobot-scheduler.pid Restart=always RestartSec=30 PrivateTmp=true [Install] WantedBy=multi-user.target
Migrating to Celery from RQ¶
Prior to migrating, you need to determine whether you have any plugins installed that run custom background tasks that still rely on the RQ worker. There are a few ways to do this. Two of them are:
- Ask your developer or administrator if there are any plugins running background tasks still using the RQ worker
- If you are savvy with code, search your code for the
@jobdecorator or for
from django_rq import job
If you're upgrading from Nautobot version 1.0.x and are NOT running plugins that use the RQ worker, all you really need to do are two things.
First, you must replace the contents of
/etc/systemd/system/nautobot-worker.service with the
systemd unit file provided just above.
Next, you must update any custom background tasks that you may have written. If you do not have any custom background tasks, then you may continue on to the next section to reload your worker service to use Celery.
To update your custom tasks, you'll need to do the following.
- Replace each import
from django_rq import jobwith
from nautobot.core.celery import nautobot_task
- Replace each decorator of
diff --git a/task_example.py b/task_example.py index f84073fb5..52baf6096 100644 --- a/task_example.py +++ b/task_example.py @@ -1,6 +1,6 @@ -from django_rq import job +from nautobot.core.celery import nautobot_task -@job("default") +@nautobot_task def example_task(*args, **kwargs): return "examples are cool!" (END)
If you are using plugins that use custom background tasks but have not yet made the change described above, you must run the RQ worker concurrently with the Celery worker until the plugin can be updated.
Failure to account for the Celery-to-RQ migration may break your custom background tasks
Concurrent Celery and RQ Nautobot Workers¶
If you must run the Celery and RQ workers concurrently, you must also configure the (deprecated) RQ worker.
Copy and paste the following into
[Unit] Description=Nautobot Request Queue Worker Documentation=https://docs.nautobot.com/projects/core/en/stable/ After=network-online.target Wants=network-online.target [Service] Type=simple Environment="NAUTOBOT_ROOT=/opt/nautobot" User=nautobot Group=nautobot WorkingDirectory=/opt/nautobot ExecStart=/opt/nautobot/bin/nautobot-server rqworker Restart=on-failure RestartSec=30 PrivateTmp=true [Install] WantedBy=multi-user.target
Because we just added new service files, you'll need to reload the systemd daemon:
Then, start the
nautobot-scheduler services and enable them to initiate at boot time:
If you are also running the RQ worker, repeat the above command for the RQ service:
If you are running the concurrent RQ worker, you must remember to enable/check/restart the
nautobot-rq-worker process as needed, oftentimes in addition to the
Verify the service¶
You can use the command
systemctl status nautobot.service to verify that the WSGI service is running:
● nautobot.service - Nautobot WSGI Service Loaded: loaded (/etc/systemd/system/nautobot.service; enabled; vendor preset: enabled) Active: active (running) since Fri 2021-03-05 22:23:33 UTC; 35min ago Docs: https://docs.nautobot.com/projects/core/en/stable/ Main PID: 6992 (nautobot-server) Tasks: 16 (limit: 9513) Memory: 221.1M CGroup: /system.slice/nautobot.service ├─6992 /opt/nautobot/bin/python3 /opt/nautobot/bin/nautobot-server start /> ├─7007 /opt/nautobot/bin/python3 /opt/nautobot/bin/nautobot-server start /> ├─7010 /opt/nautobot/bin/python3 /opt/nautobot/bin/nautobot-server start /> ├─7013 /opt/nautobot/bin/python3 /opt/nautobot/bin/nautobot-server start /> ├─7016 /opt/nautobot/bin/python3 /opt/nautobot/bin/nautobot-server start /> └─7019 /opt/nautobot/bin/python3 /opt/nautobot/bin/nautobot-server start />
If the Nautobot service fails to start, issue the command
journalctl -eu nautobot.service to check for log messages that may indicate the problem.
Once you've verified that the WSGI service and worker are up and running, move on to HTTP server setup.
Operational Error: Incorrect string value¶
When using MySQL as a database backend, if you encounter a server error along the lines of
Incorrect string value: '\\xF0\\x9F\\x92\\x80' for column, it is because you are running afoul of the legacy implementation of Unicode (aka
utf8) encoding in MySQL. This often occurs when using modern Unicode glyphs like the famous poop emoji.
Please see the configuration guide on MySQL Unicode settings for instructions on how to address this.
Please see Computed fields with fallback value that is unicode results in OperationalError (#645) for more details.
SVG images not rendered¶
When serving Nautobot directly from uWSGI on RedHat or CentOS there may be a problem rendering .svg images to include the Nautobot logo. On the RedHat based operating systems there is no file
/etc/mime.types by default, unfortunately, uWSGI looks for this file to serve static files (see Serving static files with uWSGI). To work around this copy the file
/etc/mime.types from a known good system for example an Ubuntu/Debian system or even the Nautobot container to /opt/nautobot/mime.types. Then add the following line to your
uwsgi.ini file and restart the Nautobot services:
Alternatively, host Nautobot behind Nginx as instructed in HTTP server setup.