Installing wire-server (demo) components using helm


The following will install a demo version of all the wire-server components including the databases. This setup is not recommended in production but will get you started.

Demo version means

  • easy setup - only one single machine with kubernetes is needed (make sure you have at least 4 CPU cores and 8 GB of memory available)

  • no data persistence (everything stored in memory, will be lost)

What will be installed?

  • wire-server (API) - user accounts, authentication, conversations - assets handling (images, files, …) - notifications over websocket

  • wire-webapp, a fully functioning web client (like

  • wire-account-pages, user account management (a few pages relating to e.g. password reset)

What will not be installed?


You need to have access to a kubernetes cluster, and the helm local binary on your PATH.

Type helm version, you should, if everything is configured correctly, see a result like this:

Client: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}
Server: &version.Version{SemVer:"v2.13.1", GitCommit:"618447cbf203d147601b4b9bd7f8c37a5d39fbb4", GitTreeState:"clean"}

In case kubectl version shows both Client and Server versions, but helm version does not show a Server version, you may need to run helm init. The exact version (assuming v2.X.X - at the time of writing v3 is not yet supported) matters less as long as both Client and Server versions match (or are very close).

How to start installing charts from wire

Enable the wire charts helm repository:

helm repo add wire

(You can see available helm charts by running helm search wire/. To see new versions as time passes, you may need to run helm repo update)

Great! Now you can start installing.


all commands below can also take an extra --namespace <your-namespace> if you don’t want to install into the default kubernetes namespace.

Watching changes as they happen

Open a terminal and run

kubectl get pods -w

This will block your terminal and show some things happening as you proceed through this guide. Keep this terminal open and open a second terminal.

How to install in-memory databases and external components

In your second terminal, first install databases:

helm upgrade --install databases-ephemeral wire/databases-ephemeral --wait

You should see some pods being created in your first terminal as the above command completes.

You can do the following two steps (mock aws services and demo smtp server) in parallel with the above in two more terminals, or sequentially after database-ephemeral installation has succeeded.

helm upgrade --install fake-aws wire/fake-aws --wait
helm upgrade --install smtp wire/demo-smtp --wait

How to install wire-server itself


the following makes use of overrides for helm charts. You may wish to read Overriding helm configuration settings first.*

Change back to the wire-server-deploy directory. Copy example demo values and secrets:

mkdir -p wire-server && cd wire-server
cp ../values/wire-server/demo-secrets.example.yaml secrets.yaml
cp ../values/wire-server/demo-values.example.yaml values.yaml

Or, if you are not in wire-server-deploy, download example demo values and secrets:

mkdir -p wire-server && cd wire-server
curl -sSL > secrets.yaml
curl -sSL > values.yaml

Open values.yaml and replace and other domains and subdomains with domains of your choosing. Look for the # change this comments. You can try using sed -i 's/<your-domain>/g' values.yaml.

Generate some secrets (if you are using the docker image from Installing kubernetes for a demo installation (on a single virtual machine), you should open a shell on the host system for this):

openssl rand -base64 64 | env LC_CTYPE=C tr -dc a-zA-Z0-9 | head -c 42 > restund.txt
docker run --rm /dist/zauth -m gen-keypair -i 1 > zauth.txt
  1. Add the generated secret from restund.txt to secrets.yaml under brig.secrets.turn.secret

  2. add both the public and private parts from zauth.txt to secrets.yaml under brig.secrets.zAuth

  3. Add the public key from zauth.txt also to secrets.yaml under nginz.secrets.zAuth.publicKeys

Great, now try the installation:

helm upgrade --install wire-server wire/wire-server -f values.yaml -f secrets.yaml --wait

How to set up DNS records

An installation needs 5 or 6 domain names (5 without audio/video support, 6 with audio/video support):

You need

  • two DNS names for the so-called “nginz” component of wire-server (the main REST API entry point), these are usually called nginz-https.<domain> and nginz-ssl.<domain>.

  • one DNS name for the asset store (images, audio files etc. that your users are sharing); usually assets.<domain> or s3.<domain>.

  • one DNS name for the webapp (equivalent of, i.e. the javascript app running in the browser), usually called webapp.<domain>.

  • one DNS name for the account pages (hosts some html/javascript pages for e.g. password reset), usually called account.<domain>.

  • (optional) one DNS name for team settings (to manage team membership if using PRO accounts), usually called teams.<domain>

  • (optional) one DNS name for a audio/video calling server, usually called restund01.<domain>.

If you are on the most recent charts from wire-server-deploy, these are your names:

  • nginz-https.<domain>

  • nginz-ssl.<domain>

  • webapp.<domain>

  • assets.<domain>

  • account.<domain>

  • teams.<domain>

(Yes, they all need to point to the same IP address - this is necessary for the nginx ingress to know how to do internal routing based on virtual hosting.)

You may be happy with skipping the DNS setup and just make sure that the /etc/hosts on your client machine points all the above names to the right IP address: nginz-https.<domain> nginz-ssl.<domain> assets.<domain> webapp.<domain> teams.<domain> account.<domain>

How to direct traffic to your cluster

There are a few options available. The easiest option is to use an ingress with a node port, as this works everywhere and doesn’t need a special setup.

# (assuming you're in the root directory of wire-server-deploy)
mkdir -p nginx-ingress-services && cd nginx-ingress-services
cp ../values/nginx-ingress-services/demo-secrets.example.yaml secrets.yaml
cp ../values/nginx-ingress-services/demo-values.example.yaml values.yaml

You should now have the following directory structure:

├── nginx-ingress-services
│   ├── secrets.yaml
│   └── values.yaml
└── my-wire-server
    ├── secrets.yaml
    └── values.yaml

Inside the nginx-ingress-services directory, open values.yaml and replace with a domain of your choosing. You can try using sed -i 's/<your-domain>/g' values.yaml.

Next, open secrets.yaml and add a TLS wildcard certificate and private key matching your domain. For, you need a certificate for * The easiest and cheapest option is Let’s Encrypt


Let’s Encrypt & cert-manager

As an alternative to providing your own certificate, you may want to allow for automated certificate issuing through Let’s Encrypt. For this, you have to install the cert-manager first:

helm upgrade --install -n cert-manager-ns --set 'installCRDs=true' cert-manager jetstack/cert-manager

Afterwards, you have to make some minor adjustments to the nginx-ingress-services/values.yaml you have just copied and edited. Make sure the following properties are set accordingly:

  enabled: true
  useCertManager: true

  # NOTE: You may set this to `true` when deploying the first time, just to make
  #       sure everything is order, and only to `false` before deploying again, so
  #       that a valid certificate is actually issued.
  inTestMode: false
  certmasterEmail: "ADD-VALID-ADDRESS-HERE"

Please note, in this case, you can omit the secrets.yaml file entirely.

Install the nodeport nginx ingress:

helm upgrade --install nginx-ingress-controller wire/nginx-ingress-controller --wait
helm upgrade --install nginx-ingress-services wire/nginx-ingress-services -f values.yaml -f secrets.yaml --wait

Next, we want to redirect port 443 to the port the nginx https ingress nodeport is listening on (31773), and, redirect port 80 to the nginz http port (31772) (for redirects only). To do that, you have two options:

  • Option 1: ssh into your kubernetes node, then execute:

    • iptables -t nat -A PREROUTING -p tcp --dport 443 -j REDIRECT --to-port 31773

    • iptables -t nat -A PREROUTING -p tcp --dport 80 -j REDIRECT --to-port 31772

  • Option 2: Use ansible to do that, run the iptables playbook

Trying things out

At this point, with a bit of luck, everything should be working (if not, see Troubleshooting)

Can you reach the nginz server?

curl -i https://nginz-https.<domain>/status

You should get a 200 return code

HTTP/1.1 200 OK
Content-Type: text/plain
Date: ...
Server: nginx
Content-Length: 0

Can you access the webapp? Open https://webapp.<your-domain> in your browser (Firefox/Chrome/Safari only)


Which version am I on?

There are multiple artifacts which combine to form a running wire-server deployment; these include:

  • docker images for each service

  • Kubernetes configs for each deployment (from helm charts)

  • configuration maps for each deployment (from helm charts)

If you wish to get some information regarding the code currently running on your cluster you can run the following from wire-server-deploy (if you don’t have wire-server-deploy, git clone && cd wire-server-deploy first):

./bin/ <namespace> <deployment-name (e.g. brig)>

Example run:

./ demo brig
chart_version:              wire-server-0.24.9
wire_server_commit:         8ec8b7ce2e5a184233aa9361efa86351c109c134
wire_server_deploy_commit:  01e0f261ca8163e63860f8b2af6d4ae329a32c14

Note you’ll need kubectl, git and helm installed

It will output the running docker image; the corresponding wire-server commit hash (and link) and the wire-server helm chart version which is running. This will be helpful for any support requests.

Helm install / upgrade failed

Usually, you want to run:

kubectl get pods --all-namespaces

And look for any pods that are not Running. Then you can:

kubectl --namespace <namespace> logs <name-of-pod>


kubectl --namespace <namespace> describe <name-of-pod>

to know more.

As long as nobody is using your cluster yet, you can safely delete and re-create a specific Helm release (list releases with helm list --all). Example delete the wire-server Helm release:

helm delete --purge wire-server

It doesn’t work, but my problem isn’t listed here. Help!

Feel free to open a github issue or pull request here and we’ll try to improve the documentation.