Erno Kuvaja 781da9354f Remove configs and entries for deprecated registry

Including removal of the example conf in docs.
Including glance-registry command entrypoint to eventlet.
Including rpc_controller from wsgi, changed to reject.
Not including the files devstack is depending on.

This change means that glance-registry starts and announces
that it has been removed.
This change means that when ran as wsgi app, all calls to
/rpc endpoint will be rejected.

This will allow devstack to make a graceful transition to
life without the glance registry, which was deprecated in
Newton release.

Change-Id: I7bf3284cba4c38605fb50b3c458e53f896f34086

2020-06-08 21:11:45 +01:00

8.1 KiB

Raw Blame History

Controlling Glance Servers

This section describes the ways to start, stop, and reload Glance's server programs.

Starting a server

There are two ways to start a Glance server:

Manually calling the server program
Using the glance-control server daemon wrapper program

We recommend using the second method.

Manually starting the server

The first is by directly calling the server program, passing in command-line options and a single argument for a paste.deploy configuration file to use when configuring the server application.

Note

Glance ships with an etc/ directory that contains sample paste.deploy configuration files that you can copy to a standard configuration directory and adapt for your own uses. Specifically, bind_host must be set properly.

If you do not specify a configuration file on the command line, Glance will do its best to locate a configuration file in one of the following directories, stopping at the first config file it finds:

$CWD
~/.glance
~/
/etc/glance
/etc

The filename that is searched for depends on the server application name. So, if you are starting up the API server, glance-api.conf is searched for.

If no configuration file is found, you will see an error, like:

$ glance-api
ERROR: Unable to locate any configuration file. Cannot load application glance-api

Here is an example showing how you can manually start the glance-api server in a shell.:

$ sudo glance-api --config-file glance-api.conf --debug &
jsuh@mc-ats1:~$ 2011-04-13 14:50:12    DEBUG [glance-api] ********************************************************************************
2011-04-13 14:50:12    DEBUG [glance-api] Configuration options gathered from config file:
2011-04-13 14:50:12    DEBUG [glance-api] /home/jsuh/glance-api.conf
2011-04-13 14:50:12    DEBUG [glance-api] ================================================
2011-04-13 14:50:12    DEBUG [glance-api] bind_host                      65.114.169.29
2011-04-13 14:50:12    DEBUG [glance-api] bind_port                      9292
2011-04-13 14:50:12    DEBUG [glance-api] debug                          True
2011-04-13 14:50:12    DEBUG [glance-api] default_store                  file
2011-04-13 14:50:12    DEBUG [glance-api] filesystem_store_datadir       /home/jsuh/images/
2011-04-13 14:50:12    DEBUG [glance-api] registry_host                  65.114.169.29
2011-04-13 14:50:12    DEBUG [glance-api] registry_port                  9191
2011-04-13 14:50:12    DEBUG [glance-api] ********************************************************************************
2011-04-13 14:50:12    DEBUG [routes.middleware] Initialized with method overriding = True, and path info altering = True
2011-04-13 14:50:12    DEBUG [eventlet.wsgi.server] (21354) wsgi starting up on http://65.114.169.29:9292/

$ ps aux | grep glance
root     20009  0.7  0.1  12744  9148 pts/1    S    12:47   0:00 /usr/bin/python /usr/bin/glance-api glance-api.conf --debug
jsuh     20017  0.0  0.0   3368   744 pts/1    S+   12:47   0:00 grep glance

Simply supply the configuration file as the parameter to the --config-file option (the etc/glance-api.conf sample configuration file was used in the above example) and then any other options you want to use. (--debug was used above to show some of the debugging output that the server shows when starting up. Call the server program with --help to see all available options you can specify on the command line.)

For more information on configuring the server via the paste.deploy configuration files, see the section entitled Configuring Glance servers <configuring>

Note that the server daemonizes itself by using the standard shell backgrounding indicator, &, in the previous example. For most use cases, we recommend using the glance-control server daemon wrapper for daemonizing. See below for more details on daemonization with glance-control.

Using the `glance-control` program to start the server

The second way to start up a Glance server is to use the glance-control program. glance-control is a wrapper script that allows the user to start, stop, restart, and reload the other Glance server programs in a fashion that is more conducive to automation and scripting.

Servers started via the glance-control program are always daemonized, meaning that the server program process runs in the background.

To start a Glance server with glance-control, simply call glance-control with a server and the word "start", followed by any command-line options you wish to provide. Start the server with glance-control in the following way:

$ sudo glance-control [OPTIONS] <SERVER> start [CONFPATH]

Note

You must use the sudo program to run glance-control currently, as the pid files for the server programs are written to /var/run/glance/

Here is an example that shows how to start the glance-api server with the glance-control wrapper script. :

$ sudo glance-control api start glance-api.conf
Starting glance-api with /home/jsuh/glance.conf

$ ps aux | grep glance
root     20038  4.0  0.1  12728  9116 ?        Ss   12:51   0:00 /usr/bin/python /usr/bin/glance-api /home/jsuh/glance-api.conf
jsuh     20042  0.0  0.0   3368   744 pts/1    S+   12:51   0:00 grep glance

The same configuration files are used by glance-control to start the Glance server programs, and you can specify (as the example above shows) a configuration file when starting the server.

In order for your launched glance service to be monitored for unexpected death and respawned if necessary, use the following option:

$ sudo glance-control [service] start --respawn ...

Note that this will cause glance-control itself to remain running. Also note that deliberately stopped services are not respawned, neither are rapidly bouncing services (where process death occurred within one second of the last launch).

By default, output from glance services is discarded when launched with glance-control. In order to capture such output via syslog, use the following option:

$ sudo glance-control --capture-output ...

Stopping a server

If you started a Glance server manually and did not use the & backgrounding function, simply send a terminate signal to the server process by typing Ctrl-C

If you started the Glance server using the glance-control program, you can use the glance-control program to stop it. Simply do the following:

$ sudo glance-control <SERVER> stop

as this example shows:

$ sudo glance-control api stop
Stopping glance-api  pid: 17602  signal: 15

Restarting a server

You can restart a server with the glance-control program, as demonstrated here:

$ sudo glance-control api restart etc/glance-api.conf
Stopping glance-api  pid: 17611  signal: 15
Starting glance-api with /home/jpipes/repos/glance/trunk/etc/glance-api.conf

Reloading a server

You can reload a server with the glance-control program, as demonstrated here:

$ sudo glance-control api reload
Reloading glance-api (pid 18506) with signal(1)

A reload sends a SIGHUP signal to the master process and causes new configuration settings to be picked up without any interruption to the running service (provided neither bind_host or bind_port has changed).

8.1 KiB Raw Blame History