Move RabbitMQ to a dedicated conda environment.

.github/workflows/build_and_test_on_pr.yml

@@ -53,7 +53,7 @@ jobs:
           docker exec --tty --user root $DOCKERID wait-for-services
           docker exec --tty --user aiida $DOCKERID wait-for-services
           docker exec --tty --user aiida $DOCKERID /bin/bash -l -c '/opt/conda/envs/pgsql/bin/pg_ctl -D /home/$SYSTEM_USER/.postgresql status' # Check that postgres is up.
-          docker exec --tty --user root $DOCKERID /bin/bash -l -c 'service rabbitmq-server status' # Check that rabbitmq is up.
+          docker exec --tty --user root $DOCKERID /bin/bash -l -c '/opt/conda/envs/rmq/bin/rabbitmq-server status' # Check that rabbitmq is up.
           docker exec --tty --user aiida $DOCKERID /bin/bash -l -c 'conda create -y -n test_env python=3.8' # Check that one can create a new conda environment.
           docker exec --tty --user aiida $DOCKERID /bin/bash -l -c 'conda activate test_env' # Check that new environment works.
           sudo cp tmp/.ssh/id_rsa . # Copy id_rsa file from the mounted folder.

Dockerfile

@@ -53,7 +53,6 @@ RUN apt-get update && apt-get install -y --no-install-recommends  \
     locales               \
     less                  \
     psmisc                \
-    rabbitmq-server       \
     rsync                 \
     ssh                   \
     unzip                 \
@@ -95,6 +94,9 @@ RUN cd /tmp && \
 # Install PostgreSQL in a dedicated conda environment.
 RUN conda create -c conda-forge -n pgsql postgresql=10.6 && conda clean --all -f -y
 
+# Install RabbitMQ in a dedicated conda environment.
+RUN conda create -c conda-forge -n rmq rabbitmq-server && conda clean --all -f -y
+
 # Copy the script load-singlesshagent.sh to /usr/local/bin.
 COPY bin/load-singlesshagent.sh /usr/local/bin/load-singlesshagent.sh
 

my_init.d/start-rabbitmq.sh

@@ -1,21 +1,4 @@
 #!/bin/bash
 set -em
 
-DIR_RABBITMQ="/home/${SYSTEM_USER}/.rabbitmq"
-
-mkdir -p "${DIR_RABBITMQ}"
-chown rabbitmq:rabbitmq "${DIR_RABBITMQ}"
-
-# Set base directory for RabbitMQ to persist its data. This needs to be set to a folder in the system user's home
-# directory as that is the only folder that is persisted outside of the container.
-echo MNESIA_BASE="${DIR_RABBITMQ}" >> /etc/rabbitmq/rabbitmq-env.conf
-echo LOG_BASE="${DIR_RABBITMQ}/log" >> /etc/rabbitmq/rabbitmq-env.conf
-
-# Explicitly define the node name. This is necessary because the mnesia subdirectory contains the hostname, which by
-# default is set to the value of $(hostname -s), which for docker containers, will be a random hexadecimal string. Upon
-# restart, this will be different and so the original mnesia folder with the persisted data will not be found. The
-# reason RabbitMQ is built this way is through this way it allows to run multiple nodes on a single machine each with
-# isolated mnesia directories. Since in the AiiDA setup we only need and run a single node, we can simply use localhost.
-echo NODENAME=rabbit@localhost >> /etc/rabbitmq/rabbitmq-env.conf
-
-service rabbitmq-server start
+su -c /opt/start-rabbitmq.sh ${SYSTEM_USER}

opt/start-rabbitmq.sh

@@ -0,0 +1,24 @@
+#!/bin/bash
+set -em
+
+# Activate the conda environment with RabbitMQ installed in it.
+conda activate rmq
+
+DIR_RABBITMQ="/home/${SYSTEM_USER}/.rabbitmq"
+
+mkdir -p "${DIR_RABBITMQ}"
+#chown rabbitmq:rabbitmq "${DIR_RABBITMQ}"
+
+# Set base directory for RabbitMQ to persist its data. This needs to be set to a folder in the system user's home
+# directory as that is the only folder that is persisted outside of the container.
+echo MNESIA_BASE="${DIR_RABBITMQ}" >> /opt/conda/envs/rmq/etc/rabbitmq/rabbitmq-env.conf
+echo LOG_BASE="${DIR_RABBITMQ}/log" >> /opt/conda/envs/rmq/etc/rabbitmq/rabbitmq-env.conf
+
+# Explicitly define the node name. This is necessary because the mnesia subdirectory contains the hostname, which by
+# default is set to the value of $(hostname -s), which for docker containers, will be a random hexadecimal string. Upon
+# restart, this will be different and so the original mnesia folder with the persisted data will not be found. The
+# reason RabbitMQ is built this way is through this way it allows to run multiple nodes on a single machine each with
+# isolated mnesia directories. Since in the AiiDA setup we only need and run a single node, we can simply use localhost.
+echo NODENAME=rabbit@localhost >> /opt/conda/envs/rmq/etc/rabbitmq/rabbitmq-env.conf
+
+rabbitmq-server -detached