How do I load a RabbitMQ config at startup to confirm that broker objects (queues, exchanges, bindings, users, virtual hosts, permissions and parameters) are created?
According to the RabbitMQ documentation, it can be done via load_definitions http://www.rabbitmq.com/management.html#load-definitions
But I can't figure out how to use it. Would someone mind sharing an example of how this works? I can't find any examples online.
There's two bits that the documentation leaves to be desired that were stumbling blocks for me.
Generating the definitions file
I found the easiest way to do that is to configure one RabbitMQ server how you like it and then...
Go to the management web interface
Look at the bottom of the Overview tab/page for the "Import / export definitions" heading
Click the "Download broker definitions" button in that section
Configuring RabbitMQ to look for a definitions file at startup
Put the definitions file somewhere on the filesystem that it can be read by the user that your rabbitmq daemon will be running as.
Include a block like this in the configuration file:
{rabbitmq_management, [
{listener, [...]},
{load_definitions, "/etc/rabbitmq/definitions.json"} ]},
Upon startup, those definitions should get loaded. Any errors loading them should be apparent in the logs.
Related
Currently we have to restart radiusd every time after making configuration changes. According to this man page (https://freeradius.org/radiusd/man/radmin.html) some configurations can be huped using radmin so that dynamic configuration change can be done without server restart. However, this page does not give much details on what configuration can be huped, eg. I tried to hup an eap configuration but it is not supported. Is there a check list of modules/parameters configuration that are supported by hup method?
The application installed on Tomcat (9.0.36.B.RELEASE) connects to an external source using SSL (in this case, IBM MQ). I therefore need to specify a truststore, and password for this truststore.
I got it to work adding this to the JVM_OPTS in setenv.sh:
-Djavax.net.ssl.trustStore=/opt/apps/certs/myapplication.truststore
-Djavax.net.ssl.trustStorePassword=TRUSTSTORE-PASSWORD
-Djavax.net.ssl.keyStore=/opt/apps/certs/myapplication.keystore
-Djavax.net.ssl.keyStorePassword=KEYSTORE-PASSWORD
Classic issue with this, the passwords are visible when I do ps -ef|grep java (runs on RHEL).
I have seen a couple suggestions on how to do this (like Hiding plain text password in JVM startup argumnets. " ps -ef | grep 'javax.net.ssl.keyStorePassword'"). But these change/add code or configuration in the WAR file.
I am looking for a "tomcat" solution. Is there a way to to this within the confines of Tomcat. Changes to the WAR file are difficult to implement, as the application comes from a vendor.
Note: this is not for a connector configuration in web.xml since that would only setup SSL for incoming connections. Here the application is making connections to an external system (so outbound from Tomcat's perspective).
You can add additional properties to $CATALINA_BASE/conf/catalina.properties and they will be sourced during Tomcat's startup. As you are certainly aware, system properties are global to the JVM, so there is no way to restrict this configuration to a single application only: the entire Tomcat server will be affected.
Almost all Java system properties can by set this way, with a few exceptions:
catalina.base and catalina.home (obviously),
the configuration for Tomcat logging,
the configuration for JMX and other tools that start before user code.
Attention: You must check whether the VersionLoggerListener (defined in server.xml) does not have logProps="true", otherwise the values of system properties will be logged. By default only the JVM arguments are logged.
I am trying to deploy a RavenDB instance on a Kubernetes cluster. The deployment should be fully automated, i.e. there should be no need to access the UI to configure something.
I have found plenty of documentation on how raven in a container can be configured, e.g. with command line args via RAVEN_ARGS, environment variables (e.g. RAVEN_License_Eula_Accepted), or a custom settings.json file in a mounted volume.
I have tried all the options above, and they all work, except when trying to set a license. I have tried to set either License directly as a JSON string or License.Path pointing to a license.json file mounted in a volume. Yet whenever I access the UI after deploying the container, I get a notification telling me I need to set a license.
Can anyone tell me how I can get Raven to use the license I provide via the approaches mentioned above?
Thanks
You need to bootstrap the cluster with some kind of operation for the license to be picked up. For example, create a database or call the /admin/cluster/bootstrap endpoint.
I'm troubleshooting RabbitMQ's cluster network partition events and some log messages were being dropped. The nodes run on Windows VMs. I was trying to fix the message dropping part as described here. Therefore, I added the following to my advanced.config file:
[
{lager, [
{error_logger_hwm, 1024}
]}
].
How do I verify that the configuration change was applied instead of just waiting to see if more messages are dropped or not?
[UPDATE]: On my original post I was trying to change this config in the .conf file, since that's what I'm using to configure RabbitMQ. However, the lager configuration has to be done in the advanced.config file. The advanced.config file seems to be applied even if you are using a .conf file for the basic configuration.
You can't set that value in rabbitmq.conf. The link you provide shows how to set the value in the /etc/rabbitmq/advanced.config file. Please carefully re-read this comment.
You can verify it by running this command:
rabbitmqctl eval 'application:get_env(lager, error_logger_hwm).'
Also see this article:
https://www.rabbitmq.com/configure.html#verify-configuration-effective-configuration
NOTE: the RabbitMQ team monitors the rabbitmq-users mailing list and only sometimes answers questions on StackOverflow.
I want to build a RabbitMQ cluster in my dev machine (windows).
reason is that I would like to test and study it.
Is it possible to run more than one rabbitmq instance on one machine?
I am guessing I need to:
Change the listening port
Change the appdata folder (C:\Users\MyUser\AppData\Roaming)
Change the ui plugin port so I can view all instances.
Remove the service and run from cli
Has anyone tried it?
Is there a known guide?
This thread appears to detail what you need. From the thread responses (scroll down):
The following settings are necessary to separate the node runtimes
completely, while allowing them to share the installed code base.
RABBITMQ_CONFIG_FILE /etc/rabbitmq/rabbitmq
RABBITMQ_MNESIA_BASE /var/lib/rabbitmq/mnesia
RABBITMQ_LOG_BASE /var/log/rabbitmq
RABBITMQ_ENABLED_PLUGINS_FILE /etc/rabbitmq/enabled_plugins
Now the official RabbitMQ documentation contains a section "A Cluster on a Single Machine", which describes how to run multiple Rabbit nodes on a single machine.
See https://www.rabbitmq.com/clustering.html#single-machine
One approach that I took when testing out clustering on my dev machine was to fire up several VMs. I had several for Rabbit and one for HAProxy.
While you can, as per one of the other answers, run multiple instances on a single OS instance, the multiple VM approach allowed me to mimic the intended production environment more closely.
NB. I have chosen to interpret your 'one machine' as meaning a physical machine for the purpose of this answer. If your meaning was 'virtual' then I defer to the other answer posted against this question which details how to accomplish this.
You can run multiple RabbitMQ instances on 1 machine without clustering. You just need to change the ports and the node name in rabbitmq-defaults, rabbitmq-env and config files.
For Linux:
Make sure you have Erlang installed, then download the latest version for Generic Unix and extract it.
Go to rabbit installation folder(the folder you just extracted)/sbin -> open the rabbitmq-defaults file and change SYS_PREFIX=${RABBITMQ_HOME} to create the default folder somewhere else, you should set the path to the extracted folder, for ex. /home/YOUR-USERNAME/rabbit_server-msg(this is my extracted folder).
Change the node name. Open the rabbitmq-env file -> find NODENAME=rabbit#${HOSTNAME} and change the name rabbit (ex. rabbitMSG#${HOSTNAME}). In the same file locate and change DEFAULT_NODE_PORT, by default it’s set to 5672. The default clustering port is set by ${DEFAULT_NODE_PORT} + 20000. So if you set the default node port to 5673, the clustering port will be 25673.
Install the Management Plugin. Navigate to rabbitmq-server/sbin and run “./rabbitmq-plugins enable rabbitmq_management”
The AMQP and HTTP ports need to be chnaged in the config file. Copy the file rabbitmq.config.example and paste it in the rabbit_server-msg(again this is my extracted forlder)/etc(this is the default folder created by RabbitMQ from step 1)/rabbitmq folder, rename the file to just rabbitmq.config. If you can’t find the file rabbitmq.config.example in the rabbit folder create a file named rabbitmq.config and copy the code from their GitHub.
Open the config file and uncomment the following lines:
– {tcp_listeners, [5672]} and change the port number (It is recommended to set the same number as the DEFAULT_NODE_PORT in rabbitmq-env for example 5673).
– {listener, [{port, 12345} and change the port number (This is for the management plugin).
NOTE: Be careful with the commas and brackets. Remove the comma after {tcp_listeners, [5672]}. Also after {listener, [{port, 12345} close the brackets like so ]}. Otherwise when you try to run ./rabbitmq-server it will show errors (It will show you on which line in the config file you have an error so you can fix it).
Now start up the server and log in the manager. Check the Listening ports under Ports and contexts to see if you made the changes. Do this for the other servers and you will be able to run as many servers as you want on 1 machine all with different listening ports. To start the RabbitMQ server automatically when you log in use the Startup Application program.
For Windows:
Setting up RabbitMQ for Windows is preatty much the same as on Linux.
Make sure you have Erlang installed, then download the latest version for Windows and extract it.
Go to rabbit(the folder you just extracted)/sbin -> open the rabbitmq-defaults.bat file and change RABBITMQ_BASE to create the default folder somewhere else, you should set the path to the extracted folder, for ex. C:/rabbit_server-msg(this is my extracted folder).
Change the node name. Open the rabbitmq-env.bat file -> find RABBITMQ_NODENAME=rabbit#!HOSTNAME! and change the name rabbit (ex. rabbitMSG#!HOSTNAME!). In the same file locate and change RABBITMQ_NODE_PORT, by default it’s set to 5672. To change the default clustering port locate and change RABBITMQ_DIST_PORT, by default it’s set to 25672.
Install the Management Plugin. Navigate to rabbit-server/sbin and run “rabbitmq-plugins.bat enable rabbitmq_management”.
The AMQP and HTTP ports need to be chnaged in the config file. Create a file named rabbitmq.config and copy the code from their GitHub. Put the file in your installation folder/RabbitMQ (ex. my config is in C:/rabbit_server-msg/RabbitMQ). Also in the rabbitmq-defaults.bat file check CONFIG_FILE it should be set to CONFIG_FILE=!RABBITMQ_BASE!\rabbitmq.
Open the config file and uncomment the following lines:
– {tcp_listeners, [5672]} and change the port number (It is recommended to set the same number as the RABBITMQ_NODE_PORTin rabbitmq-env for example 5673).
– {listener, [{port, 12345} and change the port number (This is for the management plugin).
NOTE: Be careful with the commas and brackets. Remove the comma after {tcp_listeners, [5672]}. Also after {listener, [{port, 12345} close the brackets like so ]}. Otherwise when you try to run rabbitmq-server.bat it will show errors (It will show you on which line in the config file you have an error so you can fix it).
Now start up the server and log in the manager. Check the Listening ports under Ports and contexts to see if you made the changes. Do this for the other servers and you will be able to run as many servers as you want on 1 machine all with different listening ports.
Source: https://lazareski.com/multiple-rabbitmq-instances-on-1-machine/