I need to run kubemq cluster on my desktop for development. I have installed kubemq docker container as described here https://docs.kubemq.io/installation/docker.html. But I can't figure out how to connect to it using kubemqctl utility? Is it possible in general? It shows only kubernetes clusters from kubectl config. And I don't see the way how to pass connection information. Thanks.
update #1
after following installation instructions, I see only those cluster which is working in kubernetes cluster and listed in my kubectl config.
Getting KubeMQ Cluster List...
Current Kubernetes cluster context connection: gke_xxxxx_us-central1-a_xxxx
NAME DESIRED RUNNING READY IMAGE AGE SERVICES
env-dev/kubemq 3 3 3 kubemq/kubemq:latest 1792h54m57s ClusterIP 10.0.2.211:8080,50000,9090,5228
when I try to switch conext it shows only that cluster again
Current Kubernetes cluster context connection: gke_xxxxx_us-central1-a_xxxx ? Select kubernetes cluster context [Use arrows to move, type to filter, ? for more help]
> gke_xxxxx_us-central1-a_xxxx
My .kubemqctl.yaml content is:
autointegrated: false
currentnamespace: kubemq
currentstatefulset: kubemq-cluster
host: localhost
grpcport: 50000
restport: 9090
apiport: 8080
issecured: false
certfile: ""
kubeconfigpath: ""
connectiontype: grpc
defaulttoken: XXXXXX-ed0c-4077-9a74-b53805ee3214
update #2
I am able to connect from my go code, to my locally running kubemq cluster
archi@eagle:~/prj/kubemq-go/examples/rpc/command$ docker container ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
3cd60e4373e4 kubemq/kubemq "./kubemq-run" 46 hours ago Up 46 hours 0.0.0.0:8080->8080/tcp, 0.0.0.0:9090->9090/tcp, 0.0.0.0:50000->50000/tcp modest_wescoff
But I cannot figure out how it is possible to connect to it using kubemq utility, because it does looking once to my kubectl config which cannot contain my local kubemq cluster, running in docker container, not kubernetes cluster. To list my kubemq clusters I use command:
kubemqctl cluster get
the output of the command is shown above (in update #1 section of my question)
update #3
As @mario said. I am able to attach to my queries channel and see all messages!
$kubemqctl queries attach userservice_channel
Adding 'userservice_channel' to attach list
[queries] [userservice_channel] kubemq-3cd60e4373e4 Server connected.
[queries] [userservice_channel] { "Kind": "request", "ID": "user_token_validate", "Channel": "userservice_channel", "ReplyChannel": "_INBOX.OYnfIQX2k7V9hTmevxHApp.Iqfr3HUw", "Metadata": "some-metadata", "Body": "\n\ufffd\u0001eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJ1c2VyIjoxLCJyb2xlIjoic3VwZXJ1c2Vy.......
I assume you've already installed kubemqctl command following this instruction but you probably haven't configured it yet.
Once you run kubemqctl help command, it will show you all available options including:
Usage:
kubemqctl [command]
Available Commands:
cluster Executes KubeMQ cluster management commands
commands Execute KubeMQ 'commands' RPC commands
config Run Kubemqctl configuration wizard command
First you need to run kubemqctl config which will start the configuration wizard, guiding you step by step through the basic configuration:
In second step you should select (using arrow keys) your Local docker container:
? Set Default KubeMQ Token (press Enter for default):
? Select KubeMQ install location: [Use arrows to move, type to filter, ? for more help]
Kubernetes cluster
MicroK8s
K3s
Minikube
Other Kubernetes distribution
> Local docker container
Other options you can probably left at their defaults.
I assume you already checked that your KubeMQ is up and running and its API is available with:
curl --location --request GET "http://localhost:8080/health" \
--header "Content-Type: application/json"
If so, you should be able to interact with it using your kubemqctl command line tool.
Please let me know if it helps.
kubemqctl config is merely a wizard tool which creates your .kubemqctl.yaml file based on which you can run kubemqctl commands against your kubemq instance. My .kubemqctl.yaml looks exactly the same as yours (only with different token).
As I already mentioned in my comment, single docker container isn't considered a cluster at all. From kubernetes perspective it is merely a container runtime and it is totally normal that kubemqctl similarly to kubectl will not list it as a cluster and won't be able to perform on it any cluster-related operation such as kubemqctl cluster get. You can easily verify it by changing your .kube/config name:
mv .kube/config .kube/config_temporarily_disabled
If I run kubemqctl cluster get now, it shows me the following output:
Getting KubeMQ Cluster List...
Error: Kubernetes Config File: Stat <path_to_the_kube_config_file>: No Such File Or Directory
which means that kubemqctl similarly to kubectl always looks for kubeconfig file when performing any cluster-related operation. This group of kubemqctl commands is simply designed only for this purpose and is not intended to be used with your kubemq instance deployed as a single docker container following this instruction.
It is not a cluster both in common sense and it is not considered as such by kubemqctl utility.
Note that you didn't even create it with:
kubemqctl cluster create -t <YOUR_KUBEMQ_TOKEN>
which is used to create a cluster in all kubernetes-based solutions described in the documentation: Kubernetes, MicroK8s and K3s but you won't find it in Docker section. Basically if you didn't create your kubemq instance using kubemqctl cluster create command, you also won't be able to list it using kubemqctl cluster get as this is not a cluster.
However you can still successfully run other kubemqctl commands and their sub-commands against your standalone kubemq instance running on a single docker container. You can actually run most of commands listed below (excluding cluster):
$ kubemqctl
Usage:
kubemqctl [command]
Available Commands:
cluster Executes KubeMQ cluster management commands
commands Execute KubeMQ 'commands' RPC commands
config Run Kubemqctl configuration wizard command
events Execute KubeMQ 'events' Pub/Sub commands
events_store Execute KubeMQ 'events_store' Pub/Sub commands
help Help about any command
queries Execute KubeMQ 'queries' RPC based commands
queues Execute KubeMQ 'queues' commands
Flags:
-h, --help help for kubemqctl
--version version for kubemqctl
Use "kubemqctl [command] --help" for more information about a command.
As I already mentioned your .kubemqctl.yaml looks totally fine. It is properly configured to run commands against your kubemq running in docker. When I run docker ps on my machine I can see properly deployed kubemq container:
c9adac88484f kubemq/kubemq "./kubemq-run" 3 hours ago Up 3 hours 0.0.0.0:8080->8080/tcp, 0.0.0.0:9090->9090/tcp, 0.0.0.0:50000->50000/tcp sleepy_ganguly
As you can see in the output above (or in the output added to your question as it is almost the same), it properly maps required ports exposed by container to machine's localhost address. You can also check it with netstat command:
$ sudo netstat -ntlp | egrep "8080|9090|50000"
tcp6 0 0 :::8080 :::* LISTEN 21431/docker-proxy
tcp6 0 0 :::50000 :::* LISTEN 21394/docker-proxy
tcp6 0 0 :::9090 :::* LISTEN 21419/docker-proxy
It's basically enough to be able to use such kubemq instance. Let's run a few example commands.
Running simple:
kubemqctl queues
will show us its help page which specifies how it should be used and gives some useful examples. Let's pick kubemqctl queues list as first:
$ kubemqctl queues list
CHANNELS:
NAME CLIENTS MESSAGES BYTES FIRST_SEQUENCE LAST_SEQUENCE
q1 1 12 2621 1 12
TOTAL CHANNELS: 1
CLIENTS:
CLIENT_ID CHANNEL ACTIVE LAST_SENT PENDING STALLED
##################################### q1 false 2 1 true
TOTAL CLIENTS: 1
Let's pick another one:
kubemqctl queues send
Again, when run without required flags and parameters, it shows us some helpful usage examples:
Error: Missing Arguments, Must Be 2 Arguments, Channel And A Message
Try:
# Send message to a queue channel channel
kubemqctl queue send q1 some-message
# Send message to a queue channel with metadata
kubemqctl queue send q1 some-message --metadata some-metadata
# Send 5 messages to a queues channel with metadata
kubemqctl queue send q1 some-message --metadata some-metadata -m 5
# Send message to a queue channel with a message expiration of 5 seconds
kubemqctl queue send q1 some-message -e 5
# Send message to a queue channel with a message delay of 5 seconds
kubemqctl queue send q1 some-message -d 5
# Send message to a queue channel with a message policy of max receive 5 times and dead-letter queue 'dead-letter'
kubemqctl queue send q1 some-message -r 5 -q dead-letter
Let's run one of them (slightly modified):
kubemqctl queue send queue2 some-message --metadata some-metadata
You should see similar output:
[Channel: queue2] [client id: ##############################] -> {id: ############################, metadata: some-metadata, body: some-message - (0), sent at: #### }
Now if you list our available queues with kubemqctl queues list command, you'll see among others (q1 was previously created by me) our recently created queue named queue2:
CHANNELS:
NAME CLIENTS MESSAGES BYTES FIRST_SEQUENCE LAST_SEQUENCE
q1 1 12 2621 1 12
queue2 0 1 232 1 1
TOTAL CHANNELS: 2
...
It was just an example to show you that it can be run against kubemq instance running on a single docker container but you can run the same way other kubemqctl commands such as commands, events or queries. kubemqctl has really good context help with many usage examples so you should find easily what you need.
I hope I was able to clear up all the ambiguities.