If you had the choice, would you rather take Salsa or Guacamole? Let me explain, why you should choose Guacamole over Salsa.

In this blog article, we want to take a look at one of the smaller Apache projects out there called Apache Guacamole. Apache Guacamole allows administrators to run a web based client tool for accessing remote applications and servers. This can include remote desktop systems, applications or terminal sessions. Users can simply access them by using their web browsers. No special client or other tools are required. From there, they can login and access all pre-configured remote connections that have been specified by an administrator.

Thereby, Guacamole supports a wide variety of protocols like VNC, RDP, and SSH. This way, users can basically access anything from remote terminal sessions to full fledged Graphical User Interfaces provided by operation systems like Debian, Ubuntu, Windows and many more.

Convert every window application to a web application

If we spin this idea further, technically every window application that isn’t designed to run as an web application can be transformed to a web application by using Apache Guacamole. We helped a customer to bring its legacy application to Kubernetes, so that other users could use their web browsers to run it. Sure, implementing the application from ground up, so that it follows the Cloud Native principles, is the preferred solution. As always though, efforts, experience and costs may exceed the available time and budget and in that cases, Apache Guacamole can provide a relatively easy way for realizing such projects.

In this blog article, I want to show you, how easy it is to run a legacy window application as a web app on Kubernetes. For this, we will use a Kubernetes cluster created by kind and create a Kubernetes Deployment to make kate – a KDE based text editor – our own web application. It’s just an example, so there might be better application to transform but this one should be fine to show you the concepts behind Apache Guacamole.

So, without further ado, let’s create our kate web application.

Preparation of Kubernetes

Before we can start, we must make sure that we have a Kubernetes cluster, that we can test on. If you already have a cluster, simply skip this section. If not, let’s spin one up by using kind.

kind is a lightweight implementation of Kubernetes that can be run on every machine. It’s written in Go and can be installed like this:

# For AMD64 / x86_64
[ $(uname -m) = x86_64 ] && curl -Lo ./kind https://kind.sigs.k8s.io/dl/v0.22.0/kind-linux-amd64
# For ARM64
[ $(uname -m) = aarch64 ] && curl -Lo ./kind https://kind.sigs.k8s.io/dl/v0.22.0/kind-linux-arm64
chmod +x ./kind
sudo mv ./kind /usr/local/bin/kind

Next, we need to install some dependencies for our cluster. This includes for example docker and kubectl.

$ sudo apt install docker.io kubernetes-client

By creating our Kubernetes Cluster with kind, we need docker because the Kubernetes cluster is running within Docker containers on your host machine. Installing kubectl allows us to access the Kubernetes after creating it.

Once we installed those packages, we can start to create our cluster now. First, we must define a cluster configuration. It defines which ports are accessible from our host machine, so that we can access our Guacamole application. Remember, the cluster itself is operated within Docker containers, so we must ensure that we can access it from our machine. For this, we define the following configuration which we save in a file called cluster.yaml:

kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
- role: control-plane
  extraPortMappings:
  - containerPort: 30000
    hostPort: 30000
    listenAddress: "127.0.0.1"
    protocol: TCP

Hereby, we basically map the container’s port 30000 to our local machine’s port 30000, so that we can easily access it later on. Keep this in mind because it will be the port that we will use with our web browser to access our kate instance.

Ultimately, this configuration is consumed by kind . With it, you can also adjust multiple other parameters of your cluster besides of just modifying the port configuration which are not mentioned here. It’s worth to take a look kate’s documentation for this.

As soon as you saved the configuration to cluster.yaml, we can now start to create our cluster:

$ sudo kind create cluster --name guacamole --config cluster.yaml
Creating cluster "guacamole" ...
 ✓ Ensuring node image (kindest/node:v1.29.2) 🖼
 ✓ Preparing nodes 📦
 ✓ Writing configuration 📜
 ✓ Starting control-plane 🕹️
 ✓ Installing CNI 🔌
 ✓ Installing StorageClass 💾
Set kubectl context to "kind-guacamole"
You can now use your cluster with:

kubectl cluster-info --context kind-guacamole

Have a question, bug, or feature request? Let us know! https://kind.sigs.k8s.io/#community 🙂

Since we don’t want to run everything in root context, let’s export the kubeconfig, so that we can use it with kubectl by using our unpriviledged user:

$ sudo kind export kubeconfig \
    --name guacamole \
    --kubeconfig $PWD/config

$ export KUBECONFIG=$PWD/config
$ sudo chown $(logname): $KUBECONFIG

By doing so, we are ready and can access our Kubernetes cluster using kubectl now. This is our baseline to start migrating our application.

Creation of the Guacamole Deployment

In order to run our application on Kubernetes, we need some sort of workload resource. Typically, you could create a Pod, Deployment, Statefulset or Daemonset to run workloads on a cluster.

Let’s create the Kubernetes Deployment for our own application. The example shown below shows the deployment’s general structure. Each container definition will have their dedicated examples afterwards to explain them in more detail.

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: web-based-kate
  name: web-based-kate
spec:
  replicas: 1
  selector:
    matchLabels:
      app: web-based-kate
  template:
    metadata:
      labels:
        app: web-based-kate
    spec:
      containers:
      # The guacamole server component that each
      # user will connect to via their browser
      - name: guacamole-server
        image: docker.io/guacamole/guacamole:1.5.4
        ...
      # The daemon that opens the connection to the
      # remote entity
      - name: guacamole-guacd
        image: docker.io/guacamole/guacd:1.5.4
        ...
      # Our own self written application that we
      # want to make accessible via the web.
      - name: web-based-kate
        image: registry.example.com/own-app/web-based-kate:0.0.1
        ...
      volumes:
        - name: guacamole-config
          secret:
            secretName: guacamole-config
        - name: guacamole-server
          emptyDir: {}
        - name: web-based-kate-home
          emptyDir: {}
        - name: web-based-kate-tmp
          emptyDir: {}

As you can see, we need three containers and some volumes for our application. The first two containers are dedicated to Apache Guacamole itself. First, it’s the server component which is the external endpoint for clients to access our web application. It serves the web server as well as the user management and configuration to run Apache Guacamole.

Next to this, there is the guacd daemon. This is the core component of Guacamole which creates the remote connections to the application based on the configuration done to the server. This daemon forwards the remote connection to the clients by making it accessible to the Guacamole server which then forwards the connection to the end user.

Finally, we have our own application. It will offer a connection endpoint to the guacd daemon using one of Guacamole’s supported protocols and provide the Graphical User Interface (GUI).

Guacamole Server

Now, let’s deep dive into each container specification. We are starting with the Guacamole server instance. This one handles the session and user management and contains the configuration which defines what remote connections are available and what are not.

- name: guacamole-server
  image: docker.io/guacamole/guacamole:1.5.4
  env:
    - name: GUACD_HOSTNAME
      value: "localhost"
    - name: GUACD_PORT
      value: "4822"
    - name: GUACAMOLE_HOME
      value: "/data/guacamole/settings"
    - name: HOME
      value: "/data/guacamole"
    - name: WEBAPP_CONTEXT
      value: ROOT
  volumeMounts:
    - name: guacamole-config
      mountPath: /data/guacamole/settings
    - name: guacamole-server
      mountPath: /data/guacamole
  ports:
    - name: http
      containerPort: 8080
  securityContext:
    allowPrivilegeEscalation: false
    privileged: false
    readOnlyRootFilesystem: true
    capabilities:
      drop: ["all"]
  resources:
    limits:
      cpu: "250m"
      memory: "256Mi"
    requests:
      cpu: "250m"
      memory: "256Mi"

Since it needs to connect to the guacd daemon, we have to provide the connection information for guacd by passing them into the container using environment variables like GUACD_HOSTNAME or GUACD_PORT. In addition, Guacamole would usually be accessible via http://<your domain>/guacamole.

This behavior however can be adjusted by modifying the WEBAPP_CONTEXT environment variable. In our case for example, we don’t want a user to type in /guacamole to access it but simply using it like this http://<your domain>/

Guacamole Guacd

Then, there is the guacd daemon.

- name: guacamole-guacd
  image: docker.io/guacamole/guacd:1.5.4
  args:
    - /bin/sh
    - -c
    - /opt/guacamole/sbin/guacd -b 127.0.0.1 -L $GUACD_LOG_LEVEL -f
  securityContext:
    allowPrivilegeEscalation: true
    privileged: false
    readOnlyRootFileSystem: true
    capabilities:
      drop: ["all"]
  resources:
    limits:
      cpu: "250m"
      memory: "512Mi"
    requests:
      cpu: "250m"
      memory: "512Mi"

It’s worth mentioning that you should modify the arguments used to start the guacd container. In the example above, we want guacd to only listen to localhost for security reasons. All containers within the same pod share the same network namespace. As a a result, they can access each other via localhost. This said, there is no need to make this service accessible to over services running outside of this pod, so we can limit it to localhost only. To achieve this, you would need to set the -b 127.0.0.1 parameter which sets the corresponding listen address. Since you need to overwrite the whole command, don’t forget to also specify the -L and -f parameter. The first parameter sets the log level and the second one set the process in the foreground.

Web Based Kate

To finish everything off, we have the kate application which we want to transform to a web application.

- name: web-based-kate
  image: registry.example.com/own-app/web-based-kate:0.0.1
  env:
    - name: VNC_SERVER_PORT
      value: "5900"
    - name: VNC_RESOLUTION_WIDTH
      value: "1280"
    - name: VNC_RESOLUTION_HEIGHT
      value: "720"
  securityContext:
    allowPrivilegeEscalation: true
    privileged: false
    readOnlyRootFileSystem: true
    capabilities:
      drop: ["all"]
  volumeMounts:
    - name: web-based-kate-home
      mountPath: /home/kate
    - name: web-based-kate-tmp
      mountPath: /tmp

Configuration of our Guacamole setup

After having the deployment in place, we need to prepare the configuration for our Guacamole setup. In order to know, what users exist and which connections should be offered, we need to provide a mapping configuration to Guacamole.

In this example, a simple user mapping is shown for demonstration purposes. It uses a static mapping defined in a XML file that is handed over to the Guacamole server. Typically, you would use other authentication methods instead like a database or LDAP.

This said however, let’s continue with our static one. For this, we simply define a Kubernetes Secret which is mounted into the Guacamole server. Hereby, it defines two configuration files. One is the so called guacamole.properties. This is Guacamole’s main configuration file. Next to this, we also define the user-mapping.xml which contains all available users and their connections.

apiVersion: v1
kind: Secret
metadata:
  name: guacamole-config
stringData:
  guacamole.properties: |
    enable-environment-properties: true
  user-mapping.xml: |
    <user-mapping>
      <authorize username="admin" password="PASSWORD" encoding="sha256">
        <connection name="web-based-kate">
          <protocol>vnc</protocol>
          <param name="hostname">localhost</param>
          <param name="port">5900</param>
        </connection>
      </authorize>
    </user-mapping>

As you can see, we only defined on specific user called admin which can use a connection called web-based-kate. In order to access the kate instance, Guacamole would use VNC as the configured protocol. To make this happen, our web application must offer a VNC Server port on the other side, so that the guacd daemon can then access it to forward the remote session to clients. Keep in mind that you need to replace the string PASSWORD to a proper sha256 sum which contains the password. The sha256 sum could look like this for example:

$ echo -n "test" | sha256sum
9f86d081884c7d659a2feaa0c55ad015a3bf4f1b2b0b822cd15d6c15b0f00a08  -

Next, the hostname parameter is referencing the corresponding VNC server of our kate container. Since we are starting our container alongside with our Guacamole containers within the same pod, the Guacamole Server as well as the guacd daemon can access this application via localhost. There is no need to set up a Kubernetes Service in front of it since only guacd will access the VNC server and forward the remote session via HTTP to clients accessing Guacamole via their web browsers. Finally, we also need to specify the VNC server port which is typically 5900 but this could be adjusted if needed.

The corresponding guacamole.properties is quite short. By enabling the enabling-environment-properties configuration parameter, we make sure that every Guacamole configuration parameter can also be set via environment variables. This way, we don’t need to modify this configuration file each and every time when we want to adjust the configuration but we only need to provide updated environment variables to the Guacamole server container.

Make Guacamole accessible

Last but not least, we must make the Guacamole server accessible for clients. Although each provided service can access each other via localhost, the same does not apply to clients trying to access Guacamole. Therefore, we must make Guacamole’s server port 8080 available to the outside world. This can be achieved by creating a Kubernetes Service of type NodePort. This service is forwarding each request from a local node port to the corresponding container that is offering the configured target port. In our case, this would be the Guacamole server container which is offering port 8080.

apiVersion: v1
kind: Service
metadata:
  name: web-based-kate
spec:
  type: NodePort
  selector:
    app: web-based-kate
  ports:
    - name: http
      protocol: TCP
      port: 8080
      targetPort: 8080
      nodePort: 30000

This specific port is then mapped to the Node’s 30000 port for which we also configured the kind cluster in such a way that it forwards its node port 30000 to the host system’s port 30000. This port is the one that we would need to use to access Guacamole with our web browsers.

Prepartion of the Application container

Before we can start to deploy our application, we need to prepare our kate container. For this, we simply create a Debian container that is running kate. Keep in mind that you would typically use lightweight base images like alpine to run applications like this. For this demonstration however, we use the Debian images since it is easier to spin it up but in general you only need a small friction of the functionality that is provided by this base image. Moreover – from an security point of view – you want to keep your images small to minimize the attack surface and make sure it is easier to maintain. For now however, we will continue with the Debian image.

In the example below, you can see a Dockerfile for the kate container.

FROM debian:12

# Install all required packages
RUN apt update && \
    apt install -y x11vnc xvfb kate

# Add user for kate
RUN adduser kate --system --home /home/kate -uid 999

# Copy our entrypoint in the container
COPY entrypoint.sh /opt

USER 999
ENTRYPOINT [ "/opt/entrypoint.sh" ]

Here you see that we create a dedicated user called kate (User ID 999) for which we also create a home directory. This home directory is used for all files that kate is creating during runtime. Since we set the readOnlyRootFilesystem to true, we must make sure that we mount some sort of writable volume (e.g EmptyDir) to kate’s home directory. Otherwise, kate wouldn’t be able to write any runtime data then.

Moreover, we have to install the following three packages:

• x11vnc
• xvfb
• kate

These are the only packages we need for our container. In addition, we also need to create an entrypoint script to start the application and prepare the container accordingly. This entrypoint script creates the configuration for kate, starts it in a virtual display by using xvfb-run and provides this virtual display to end users by using the VNC server via x11vnc. In the meantime, xdrrinfo is used to check if the virtual display came up successfully after starting kate. If it takes to long, the entrypoint script will fail by returning the exit code 1.

By doing this, we ensure that the container is not stuck in an infinite loop during a failure and let Kubernetes restart the container whenever it couldn’t start the application successfully. Furthermore, it is important to check if the virtual display came up prior of handing it over to the VNC server because the VNC server would crash if the virtual display is not up and running since it needs something to share. On the other hand though, our container will be killed whenever kate is terminated because it would also terminate the virtual display and in the end it would then also terminate the VNC server which let’s the container exit, too. This way, we don’t need take care of it by our own.

#!/bin/bash

set -e

# If no resolution is provided
if [ -z $VNC_RESOLUTION_WIDTH ]; then
  VNC_RESOLUTION_WIDTH=1920
fi

if [ -z $VNC_RESOLUTION_HEIGHT ]; then
  VNC_RESOLUTION_HEIGHT=1080
fi

# If no server port is provided
if [ -z $VNC_SERVER_PORT ]; then
  VNC_SERVER_PORT=5900
fi

# Prepare configuration for kate
mkdir -p $HOME/.local/share/kate
echo "[MainWindow0]
"$VNC_RESOLUTION_WIDTH"x"$VNC_RESOLUTION_HEIGHT" screen: Height=$VNC_RESOLUTION_HEIGHT
"$VNC_RESOLUTION_WIDTH"x"$VNC_RESOLUTION_HEIGHT" screen: Width=$VNC_RESOLUTION_WIDTH
"$VNC_RESOLUTION_WIDTH"x"$VNC_RESOLUTION_HEIGHT" screen: XPosition=0
"$VNC_RESOLUTION_WIDTH"x"$VNC_RESOLUTION_HEIGHT" screen: YPosition=0
Active ViewSpace=0
Kate-MDI-Sidebar-Visible=false" > $HOME/.local/share/kate/anonymous.katesession

# We need to define an XAuthority file
export XAUTHORITY=$HOME/.Xauthority

# Define execution command
APPLICATION_CMD="kate"

# Let's start our application in a virtual display
xvfb-run \
  -n 99 \
  -s ':99 -screen 0 '$VNC_RESOLUTION_WIDTH'x'$VNC_RESOLUTION_HEIGHT'x16' \
  -f $XAUTHORITY \
  $APPLICATION_CMD &

# Let's wait until the virtual display is initalize before
# we proceed. But don't wait infinitely.
TIMEOUT=10
while ! (xdriinfo -display :99 nscreens); do 
  sleep 1
  let TIMEOUT-=1
done

# Now, let's make the virtual display accessible by
# exposing it via the VNC Server that is listening on
# localhost and the specified port (e.g. 5900)
x11vnc \
  -display :99 \
  -nopw \
  -localhost \
  -rfbport $VNC_SERVER_PORT \
  -forever

After preparing those files, we can now create our image and import it to our Kubernetes cluster by using the following commands:

# Do not forget to give your entrypoint script
# the proper permissions do be executed
$ chmod +x entrypoint.sh

# Next, build the image and import it into kind,
# so that it can be used from within the clusters.
$ sudo docker build -t registry.example.com/own-app/web-based-kate:0.0.1 .
$ sudo kind load -n guacamole docker-image registry.example.com/own-app/web-based-kate:0.0.1

The image will be imported to kind, so that every workload resource operated in our kind cluster can access it. If you use some other Kubernetes cluster, you would need to upload this to a registry that your cluster can pull images from.

Finally, we can also apply our previously created Kubernetes manifests to the cluster. Let’s say we saved everything to one file called kuberentes.yaml. Then, you can simply apply it like this:

$ kubectl apply -f kubernetes.yaml
deployment.apps/web-based-kate configured
secret/guacamole-config configured
service/web-based-kate unchanged

This way, a Kubernetes Deployment, Secret and Service is created which ultimately creates a Kubernetes Pod which we can access afterwards.

$ kubectl get pod
NAME                              READY   STATUS    RESTARTS   AGE
web-based-kate-7894778fb6-qwp4z   3/3     Running   0          10m

Verification of our Deployment

Now, it’s money time! After preparing everything, we should be able to access our web based kate application by using our web browser. As mentioned earlier, we configured kind in such a way that we can access our application by using our local port 30000. Every request to this port is forwarded to the kind control plane node from where it is picked up by the Kubernetes Service of type NodePort. This one is then forwarding all requests to our designated Guacamole server container which is offering the web server for accessing remote application’s via Guacamole.

If everything works out, you should be able to see the the following login screen:

After successfully login in, the remote connection is established and you should be able to see the welcome screen from kate:

If you click on New, you can create a new text file:

Those text files can even be saved but keep in mind that they will only exist as long as our Kubernetes Pod exists. Once it gets deleted, the corresponding EmptyDir, that we mounted into our kate container, gets deleted as well and all files in it are lost. Moreover, the container is set to read-only meaning that a user can only write files to the volumes (e.g. EmptyDir) that we mounted to our container.

Conclusion

After seeing that it’s relatively easy to convert every application to a web based one by using Apache Guacamole, there is only one major question left…

What do you prefer the most. Salsa or Guacamole?

Integrating Proxmox Backup Server into Proxmox Clusters

Proxmox Backup Server

In today’s digital landscape, where data reigns supreme, ensuring its security and integrity is paramount for businesses of all sizes. Enter Proxmox Backup Server, a robust solution poised to revolutionize data protection strategies with its unparalleled features and open-source nature.

At its core, Proxmox Backup Server is a comprehensive backup solution designed to safeguard critical data and applications effortlessly in virtualized environments based on Proxmox VE. Unlike traditional backup methods, Proxmox Backup Server offers a streamlined approach, simplifying the complexities associated with data backup and recovery.

One of the standout features of Proxmox Backup Server is its seamless integration with Proxmox Virtual Environment (PVE), creating a cohesive ecosystem for managing virtualized environments. This integration allows for efficient backup and restoration of Linux containers and virtual machines, ensuring minimal downtime and maximum productivity. Without the need of any backup clients on each container or virtual machine, this solution still offers the back up and restore the entire system but also single files directly from the filesystem.

Proxmox Backup Server provides a user friendly interface, making it accessible to both seasoned IT professionals and newcomers alike. With its intuitive design, users can easily configure backup tasks, monitor progress, and retrieve data with just a few clicks, eliminating the need for extensive training or technical expertise.

Data security is a top priority for businesses across industries and Proxmox Backup Server delivers on this front. Bundled with solutions like ZFS it also brings in all the enterprise filesystem features like encryption at rest, encryption at transition, checksums, snapshots, deduplication and compression but also integrating iSCSI or NFS storage from enterprise storage solutions like from NetApp can be used.

Another notable aspect of Proxmox Backup Server is its cost effectiveness. As an open-source solution, it eliminates the financial barriers (also in addition with the Proxmox VE solutions) associated with proprietary backup software.

Integrating Proxmox Backup Server into Proxmox Clusters

General

This guide expects you to have already at least one Proxmox VE system up and running and also a system where a basic installation of Proxmox Backup Server has been performed. Within this example, the Proxmox Backup Server is installed on a single disk, where the datastore gets attached to an additional block device holding the backups. Proxmox VE and Proxmox Backup Server instances must not be in the same network but must be reachable for each other. The integration requires administrative access to the datacenter of the Proxmox VE instance(s) and the Backup Server.

Prerequisites

• Proxmox VE (including the datacenter).
• Proxmox Backup Server (basic installation).
• Administrative access to all systems.
• Network reachability.
• Storage device holding the backups (in this case a dedicated block storage device).

Administration: Proxmox Backup Server

Like the Proxmox VE environment, the Proxmox Backup Server comes along with a very intuitive web frontend. Unlike the web frontend of  Proxmox VE, which runs on tcp/8006, the Proxmox Backup Server can be reached on tcp/8007. Therefore, all next tasks will be done on https://<IP-PROXMOX-BACKUP-SERVER>:8007.

After logging in to the web frontend, the dashboard overview welcomes the user.

Adding Datastore / Managing Storage

The initial and major tasks relies in managing the storage and adding a usable datastore for the virtualization environment holding the backup data. Therefore, we switch to the Administration chapter and click on Storage / Disks. This provides an overview of the available Devices on the Proxmox Backup Server. As already being said, this example uses a dedicated block storage device which will be used with ZFS to benefit from checksums, deduplication, compression which of course can also be used in addition with multiple disks (so called raidz-levels) or with other solutions like folder or NFS shares. Coming back to our example, we can see the empty /dev/sdb device which will be used to store all backup files.

By clicking on ZFS in the top menu bar, a ZFS trunk can be created as a datastore. Within this survey, a name, the raid level, compression and the devices to use must be defined. As already mentioned, we can attach multiple disks and define a desired raid level. The given example only consists of a single disk, which will be defined here. Compression is optional, but using LZ4 as a compression is recommended. As a lossless data compression algorithm, LZ4 aims to provide a good trade off between speed and compression ratio which is very transparent on today’s system.

Ensure to check Add as Datastore option (default) will create the given name directly as a usable datastore. In our example this will be backup01.

Keep in mind, that this part is not needed when using a NFS share. Also do not use this in addition with hardware RAID controllers.

Adding User for Backup

In a next step, a dedicated user will be created that will be used for the datastore permissions and for the Proxmox VE instances for authentication and authorization. This allows even complex setups with different datastores, different users including different access levels (e.g., reading, writing, auditing,…) on different clusters and instances. To keep it simple for demonstrations, just a single user for everything will be used.

A new user is configured by selecting Configuration, Access Control and User Management in the left menu. There, a new user can be created by simply defining a name and a password. The default realm should stay on the default for the Proxmox Backup authentication provider. Depending on the complexity of the used name schema, you may also create reasonable users. In the given example, the user is called dc01cluster22backup01.

Adding Permission of User for Datastore

Mentioning already the possibility to create complex setups regarding authentication and authorization, the datastore must be linked to at least a single user that can access it. Therefore, we go back to the Datastore and select the previously created backup01 datastore. In the top menu bar, the permissions can be created and adjusted in the Permissions chapter. Initially, a new one will be created now. Within the following survey the datastore or path, the user and the role must be defined:

Path: /datastore/backup01
User: dc01cluster22backup01@pbs
Role: DatastoreAdmin
Propagate: True

To provide a short overview of the possible roles, this will be shortly mentioned without any further explanation:

• Admin
• Audit
• DatastoreAdmin
• DatastoreAudit
• DatastoreBackup
• DatastorePowerUser
• DatastoreReader

Administration: Proxmox VE

The integration of the backup datastore will be performed from the Proxmox VE instances via the Datacenter. As a result, the Proxmox VE web frontend will now be used for further administrative actions. The Proxmox VE web frontend runs on tcp/8006, Therefore, all next tasks will be done on https://<IP-PROXMOX-VE-SERVER>:8006.

Adding Storage

Integrating the Proxmox Backup Server works the same way like managing and adding a shared storage to a Proxmox datacenter.

In the left menu we choose the active datacenter and select the Storage options. There, we can find all natively support storage options like (NFS, SMB/CIFS, iSCSI, ZFS, GlusterFS,…) of Proxmox and finally select the Proxmox Backup Server as a dedicated item.

Afterwards, the details for adding this datastore to the datacenter must be inserted. The following options need to be defined:

ID: backup22
Server: <FQDN-OR-IP-OF-BACKUP-SERVER>
Username: dc01cluster22backup01@pbs
Password: <THE-PASSWORD-OF-THE-USER>
Enable: True
Datastore: backup01
Fingerprint: <SYSTEM-FINGERPRINT-OF-BACKUP-SERVER>

Optionally, also the Backup Retention and Encryption can be configured before adding the new backup datastore. While the backup retention can also be configured on the Proxmox Backup Server (which is recommended), enabling the encryption should be considered. Selecting an d activating the encryption is easily done by simply setting it to Auto-generate a client encryption key. Depending on your previous setup, also an already present key can be uploaded and used.

After adding this backup datastore to the datacenter, this can immediately be used for backup and the integration is finalized.

Conclusion

Proxmox provides with the Proxmox Backup Server an enterprise backup solution, for backing up Linux containers and virtual machines. Supporting features like incremental and fully deduplicated backups by using the benefits of different open-source solutions, in addition with strong encryption and data integrity this solution is a prove that open-source software can compete with closed-source enterprise software. Together with Proxmox VE, enterprise like virtualization environments can be created and managed without missing the typical enterprise feature set. Proxmox VE and the Proxmox Backup Server can also be used in addition to storage appliances from vendors like NetApp, by directly use iSCSI or NFS.

Providing this simple example, there are of course much more complex scenarios which can be created and also should be considered. We are happy to provide you more information and to assist you creating such setups. We also provide help for migrating from other products to Proxmox VE setups. Feel free to contact us at any time for more information.

Migrating VMs from VMware ESXi to Proxmox

In response to Broadcom’s recent alterations in VMware’s subscription model, an increasing number of enterprises are reevaluating their virtualization strategies. With heightened concerns over licensing costs and accessibility to features, businesses are turning towards open source solutions for greater flexibility and cost-effectiveness. Proxmox, in particular, has garnered significant attention as a viable alternative. Renowned for its robust feature set and open architecture, Proxmox offers a compelling platform for organizations seeking to mitigate the impact of proprietary licensing models while retaining comprehensive virtualization capabilities. This trend underscores a broader industry shift towards embracing open-source technologies as viable alternatives in the virtualization landscape. Just to mention, Proxmox is widely known as a viable alternative to VMware ESXi but there are also other options available, such as bhyve which we also covered in one of our blog posts.

Benefits of Opensource Solutions

In the dynamic landscape of modern business, the choice to adopt open source solutions for virtualization presents a strategic advantage for enterprises. With platforms like KVM, Xen and even LXC containers, organizations can capitalize on the absence of license fees, unlocking significant cost savings and redirecting resources towards innovation and growth. This financial flexibility empowers companies to make strategic investments in their IT infrastructure without the burden of proprietary licensing costs. Moreover, open source virtualization promotes collaboration and transparency, allowing businesses to tailor their environments to suit their unique needs and seamlessly integrate with existing systems. Through community-driven development and robust support networks, enterprises gain access to a wealth of expertise and resources, ensuring the reliability, security, and scalability of their virtualized infrastructure. Embracing open source virtualization not only delivers tangible financial benefits but also equips organizations with the agility and adaptability needed to thrive in an ever-evolving digital landscape.

Migrating a VM

Prerequisites

To ensure a smooth migration process from VMware ESXi to Proxmox, several key steps must be taken. First, SSH access must be enabled on both the VMware ESXi host and the Proxmox host, allowing for remote management and administration. Additionally, it’s crucial to have access to both systems, facilitating the migration process. Furthermore, establishing SSH connectivity between VMware ESXi and Proxmox is essential for seamless communication between the two platforms. This ensures efficient data transfer and management during migration. Moreover, it’s imperative to configure the Proxmox system or cluster in a manner similar to the ESXi setup, especially concerning networking configurations. This includes ensuring compatibility with VLANs or VXLANs for more complex setups. Additionally, both systems should either run on local storage or have access to shared storage, such as NFS, to facilitate the transfer of virtual machine data. Lastly, before initiating the migration, it’s essential to verify that the Proxmox system has sufficient available space to accommodate the imported virtual machine, ensuring a successful transition without storage constraints.

Activate SSH on ESXi

The SSH server must be activated in order to copy the content from the ESXi system to the new location on the Proxmox server. The virtual machine will later be copied from the Proxmox server. Therefore, it is necessary that the Proxmox system can establish an SSH connection on tcp/22 to the ESXi system:

• Log in to the VMware ESXi host.
• Navigate to Configuration > Security Profile.
• Enable SSH under Services.

Find Source Information about VM on ESXi

One of the challenging matters in finding the location of the virtual machine holding the virtual machine disk. The path can be found within the web UI of the ESXi system:

• Locate the ESXi node that runs the Virtual Machine that should be migrated
• Identify the virtual machine to be migrated (e.g., pgsql07.gyptazy.ch).
• Obtain the location of the virtual disk (VMDK) associated with the VM from the configuration panel.
• The VM location path should be shown (e.g., /vmfs/volumes/137b4261-68e88bae-0000-000000000000/pgsql07.gyptazy.ch).
• Stop and shutdown the VM.

Create a New Empty VM on Proxmox

• Create a new empty VM in Proxmox.
• Assign the same resources like in the ESXi setup.
• Set the network type to VMware vmxnet3.
  • Ensure the needed network resources (e.g., VLAN, VXLAN) are properly configured.
• Set the SCSCI controller for the disk to VMware PVSCSI.
  • Do not create a new disk (this will be imported later from the ESXi source).
• Each VM gets an ID assigned by Proxmox (note it down, it will be needed later).

Copy VM from ESXi to Proxmox

The content of the virtual machine (VM) will be transferred from the ESXi to the Proxmox system using the open source tool rsync for efficient synchronization and copying. Therefore, the following commands need to be executed from the Proxmox system, where we create a temporary directory to store the VM’s content:

mkdir /tmp/migration_pgsql07.gyptazy.ch
cd /tmp/migration_pgsql07.gyptazy.ch
rsync -avP root@esx02-test.gyptazy.ch:/vmfs/volumes/137b4261-68e88bae-0000-000000000000/pgsq07.gyptazy.ch/* .

Depending on the file size of them virtual machine and the network connectivity this process may take some time.

Import VM in Proxmox

Afterwards, the disk is imported using the qm utility, defining the VM ID (which got created during the VM creation process), along with specifying the disk name (which has been copied over) and the destination data storage on the Proxmox system where the VM disk should be stored:

qm disk import 119 pgsql07.gyptazy.ch.vmdk local-lvm

Depending on the creation format of the VM or the exporting format there may be multiple disk files which may also be suffixed by _flat. This procedure needs to be repeated by all available disks.

Starting the VM

In the final step, all settings, resources, definitions and customizations of the system should be thoroughly reviewed. One validated, the VM can be launched, ensuring that all components are correctly configured for operation within the Proxmox environment.

Conclusion

This article only covers one of many possible methods for migrations in simple, standalone setups. In more complex environments involving multiple host nodes and different storage systems like fibre channel or network storage, there are significant differences and additional considerations. Additionally, there may be specific requirements regarding availability and Service Level Agreements (SLAs) to be concern. This may be very specific for each environment. Feel free to contact us for personalized guidance on your specific migration needs at any time. We are also pleased to offer our support in related areas in open source such as virtualization (e.g., OpenStack, VirtualBox) and topics pertaining to cloud migrations.

Addendum

On the 27th of March, Proxmox released their new import wizard (pve-esxi-import-tools) which makes migrations from VMware ESXi instances to a Proxmox environment much easier. Within an upcoming blog post we will provide more information about the new tooling and cases where this might be more useful but also covering the corner cases where the new import wizard cannot be used.

SQLreduce: Reduce verbose SQL queries to minimal examples

Developers often face very large SQL queries that raise some errors. SQLreduce is a tool to reduce that complexity to a minimal query.

SQLsmith generates random SQL queries

SQLsmith is a tool that generates random SQL queries and runs them against a PostgreSQL server (and other DBMS types). The idea is that by fuzz-testing the query parser and executor, corner-case bugs can be found that would otherwise go unnoticed in manual testing or with the fixed set of test cases in PostgreSQL’s regression test suite. It has proven to be an effective tool with over 100 bugs found in different areas in the PostgreSQL server and other products since 2015, including security bugs, ranging from executor bugs to segfaults in type and index method implementations. For example, in 2018, SQLsmith found that the following query triggered a segfault in PostgreSQL:

select
  case when pg_catalog.lastval() < pg_catalog.pg_stat_get_bgwriter_maxwritten_clean() then case when pg_catalog.circle_sub_pt(
          cast(cast(null as circle) as circle),
          cast((select location from public.emp limit 1 offset 13)
             as point)) ~ cast(nullif(case when cast(null as box) &> (select boxcol from public.brintest limit 1 offset 2)
                 then (select f1 from public.circle_tbl limit 1 offset 4)
               else (select f1 from public.circle_tbl limit 1 offset 4)
               end,
          case when (select pg_catalog.max(class) from public.f_star)
                 ~~ ref_0.c then cast(null as circle) else cast(null as circle) end
            ) as circle) then ref_0.a else ref_0.a end
       else case when pg_catalog.circle_sub_pt(
          cast(cast(null as circle) as circle),
          cast((select location from public.emp limit 1 offset 13)
             as point)) ~ cast(nullif(case when cast(null as box) &> (select boxcol from public.brintest limit 1 offset 2)
                 then (select f1 from public.circle_tbl limit 1 offset 4)
               else (select f1 from public.circle_tbl limit 1 offset 4)
               end,
          case when (select pg_catalog.max(class) from public.f_star)
                 ~~ ref_0.c then cast(null as circle) else cast(null as circle) end
            ) as circle) then ref_0.a else ref_0.a end
       end as c0,
  case when (select intervalcol from public.brintest limit 1 offset 1)
         >= cast(null as "interval") then case when ((select pg_catalog.max(roomno) from public.room)
             !~~ ref_0.c)
        and (cast(null as xid) <> 100) then ref_0.b else ref_0.b end
       else case when ((select pg_catalog.max(roomno) from public.room)
             !~~ ref_0.c)
        and (cast(null as xid) <> 100) then ref_0.b else ref_0.b end
       end as c1,
  ref_0.a as c2,
  (select a from public.idxpart1 limit 1 offset 5) as c3,
  ref_0.b as c4,
    pg_catalog.stddev(
      cast((select pg_catalog.sum(float4col) from public.brintest)
         as float4)) over (partition by ref_0.a,ref_0.b,ref_0.c order by ref_0.b) as c5,
  cast(nullif(ref_0.b, ref_0.a) as int4) as c6, ref_0.b as c7, ref_0.c as c8
from
  public.mlparted3 as ref_0
where true;

However, just like in this 40-line, 2.2kB example, the random queries generated by SQLsmith that trigger some error are most often very large and contain a lot of noise that does not contribute to the error. So far, manual inspection of the query and tedious editing was required to reduce the example to a minimal reproducer that developers can use to fix the problem.

Reduce complexity with SQLreduce

This issue is solved by SQLreduce. SQLreduce takes as input an arbitrary SQL query which is then run against a PostgreSQL server. Various simplification steps are applied, checking after each step that the simplified query still triggers the same error from PostgreSQL. The end result is a SQL query with minimal complexity.

SQLreduce is effective at reducing the queries from original error reports from SQLsmith to queries that match manually-reduced queries. For example, SQLreduce can effectively reduce the above monster query to just this:

SELECT pg_catalog.stddev(NULL) OVER () AS c5 FROM public.mlparted3 AS ref_0

Note that SQLreduce does not try to derive a query that is semantically identical to the original, or produces the same query result – the input is assumed to be faulty, and we are looking for the minimal query that produces the same error message from PostgreSQL when run against a database. If the input query happens to produce no error, the minimal query output by SQLreduce will just be SELECT.

How it works

We’ll use a simpler query to demonstrate how SQLreduce works and which steps are taken to remove noise from the input. The query is bogus and contains a bit of clutter that we want to remove:

$ psql -c 'select pg_database.reltuples / 1000 from pg_database, pg_class where 0 < pg_database.reltuples / 1000 order by 1 desc limit 10'
ERROR:  column pg_database.reltuples does not exist

Let’s pass the query to SQLreduce:

$ sqlreduce 'select pg_database.reltuples / 1000 from pg_database, pg_class where 0 < pg_database.reltuples / 1000 order by 1 desc limit 10'

SQLreduce starts by parsing the input using pglast and libpg_query which expose the original PostgreSQL parser as a library with Python bindings. The result is a parse tree that is the basis for the next steps. The parse tree looks like this:

selectStmt
├── targetList
│   └── /
│       ├── pg_database.reltuples
│       └── 1000
├── fromClause
│   ├── pg_database
│   └── pg_class
├── whereClause
│   └── <
│       ├── 0
│       └── /
│           ├── pg_database.reltuples
│           └── 1000
├── orderClause
│   └── 1
└── limitCount
    └── 10

Pglast also contains a query renderer that can render back the parse tree as SQL, shown as the regenerated query below. The input query is run against PostgreSQL to determine the result, in this case ERROR: column pg_database.reltuples does not exist.

Input query: select pg_database.reltuples / 1000 from pg_database, pg_class where 0 < pg_database.reltuples / 1000 order by 1 desc limit 10
Regenerated: SELECT pg_database.reltuples / 1000 FROM pg_database, pg_class WHERE 0 < ((pg_database.reltuples / 1000)) ORDER BY 1 DESC LIMIT 10
Query returns: ✔ ERROR:  column pg_database.reltuples does not exist

SQLreduce works by deriving new parse trees that are structurally simpler, generating SQL from that, and run these queries against the database. The first simplification steps work on the top level node, where SQLreduce tries to remove whole subtrees to quickly find a result. The first reduction tried is to remove LIMIT 10:

SELECT pg_database.reltuples / 1000 FROM pg_database, pg_class WHERE 0 < ((pg_database.reltuples / 1000)) ORDER BY 1 DESC ✔

The query result is still ERROR: column pg_database.reltuples does not exist, indicated by a ✔ check mark. Next, ORDER BY 1 is removed, again successfully:

SELECT pg_database.reltuples / 1000 FROM pg_database, pg_class WHERE 0 < ((pg_database.reltuples / 1000)) ✔

Now the entire target list is removed:

SELECT FROM pg_database, pg_class WHERE 0 < ((pg_database.reltuples / 1000)) ✔

This shorter query is still equivalent to the original regarding the error message returned when it is run against the database. Now the first unsuccessful reduction step is tried, removing the entire FROM clause:

SELECT WHERE 0 < ((pg_database.reltuples / 1000)) ✘ ERROR:  missing FROM-clause entry for table "pg_database"

That query is also faulty, but triggers a different error message, so the previous parse tree is kept for the next steps. Again a whole subtree is removed, now the WHERE clause:

SELECT FROM pg_database, pg_class ✘ no error

We have now reduced the input query so much that it doesn’t error out any more. The previous parse tree is still kept which now looks like this:

selectStmt
├── fromClause
│   ├── pg_database
│   └── pg_class
└── whereClause
    └── <
        ├── 0
        └── /
            ├── pg_database.reltuples
            └── 1000

Now SQLreduce starts digging into the tree. There are several entries in the FROM clause, so it tries to shorten the list. First, pg_database is removed, but that doesn’t work, so pg_class is removed:

SELECT FROM pg_class WHERE 0 < ((pg_database.reltuples / 1000)) ✘ ERROR:  missing FROM-clause entry for table "pg_database"
SELECT FROM pg_database WHERE 0 < ((pg_database.reltuples / 1000)) ✔

Since we have found a new minimal query, recursion restarts at top-level with another try to remove the WHERE clause. Since that doesn’t work, it tries to replace the expression with NULL, but that doesn’t work either.

SELECT FROM pg_database ✘ no error
SELECT FROM pg_database WHERE NULL ✘ no error

Now a new kind of step is tried: expression pull-up. We descend into WHERE clause, where we replace A < B first by A and then by B.

SELECT FROM pg_database WHERE 0 ✘ ERROR:  argument of WHERE must be type boolean, not type integer
SELECT FROM pg_database WHERE pg_database.reltuples / 1000 ✔
SELECT WHERE pg_database.reltuples / 1000 ✘ ERROR:  missing FROM-clause entry for table "pg_database"

The first try did not work, but the second one did. Since we simplified the query, we restart at top-level to check if the FROM clause can be removed, but it is still required.

From A / B, we can again pull up A:

SELECT FROM pg_database WHERE pg_database.reltuples ✔
SELECT WHERE pg_database.reltuples ✘ ERROR:  missing FROM-clause entry for table "pg_database"

SQLreduce has found the minimal query that still raises ERROR: column pg_database.reltuples does not exist with this parse tree:

selectStmt
├── fromClause
│   └── pg_database
└── whereClause
    └── pg_database.reltuples

At the end of the run, the query is printed along with some statistics:

Minimal query yielding the same error:
SELECT FROM pg_database WHERE pg_database.reltuples

Pretty-printed minimal query:
SELECT
FROM pg_database
WHERE pg_database.reltuples

Seen: 15 items, 915 Bytes
Iterations: 19
Runtime: 0.107 s, 139.7 q/s

This minimal query can now be inspected to fix the bug in PostgreSQL or in the application.

About credativ

The credativ GmbH is a manufacturer-independent consulting and service company located in Moenchengladbach, Germany. With over 22+ years of development and service experience in the open source space, credativ GmbH can assist you with unparalleled and individually customizable support. We are here to help and assist you in all your open source infrastructure needs.

Since the successful merger with Instaclustr in 2021, credativ GmbH has been the European headquarters of the Instaclustr Group, which helps organizations deliver applications at scale through its managed platform for open source technologies such as Apache Cassandra®, Apache Kafka®, Apache Spark™, Redis™, OpenSearch®, PostgreSQL®, and Cadence.
Instaclustr combines a complete data infrastructure environment with hands-on technology expertise to ensure ongoing performance and optimization. By removing the infrastructure complexity, we enable companies to focus internal development and operational resources on building cutting edge customer-facing applications at lower cost. Instaclustr customers include some of the largest and most innovative Fortune 500 companies.

Patroni is a clustering solution for PostgreSQL^® that is getting more and more popular in the cloud and Kubernetes sector due to its operator pattern and integration with Etcd or Consul. Some time ago we wrote a blog post about the integration of Patroni into Debian. Recently, the vip-manager project which is closely related to Patroni has been uploaded to Debian by us. We will present vip-manager and how we integrated it into Debian in the following.

To recap, Patroni uses a distributed consensus store (DCS) for leader-election and failover. The current cluster leader periodically updates its leader-key in the DCS. As soon the key cannot be updated by Patroni for whatever reason it becomes stale. A new leader election is then initiated among the remaining cluster nodes.

PostgreSQL Client-Solutions for High-Availability

From the user’s point of view it needs to be ensured that the application is always connected to the leader, as no write transactions are possible on the read-only standbys. Conventional high-availability solutions like Pacemaker utilize virtual IPs (VIPs) that are moved to the primary node in the case of a failover.

For Patroni, such a mechanism did not exist so far. Usually, HAProxy (or a similar solution) is used which does periodic health-checks on each node’s Patroni REST-API and routes the client requests to the current leader.

An alternative is client-based failover (which is available since PostgreSQL 10), where all cluster members are configured in the client connection string. After a connection failure the client tries each remaining cluster member in turn until it reaches a new primary.

vip-manager

A new and comfortable approach to client failover is vip-manager. It is a service written in Go that gets started on all cluster nodes and connects to the DCS. If the local node owns the leader-key, vip-manager starts the configured VIP. In case of a failover, vip-manager removes the VIP on the old leader and the corresponding service on the new leader starts it there. The clients are configured for the VIP and will always connect to the cluster leader.

Debian-Integration of vip-manager

For Debian, the pg_createconfig_patroni program from the Patroni package has been adapted so that it can now create a vip-manager configuration:

pg_createconfig_patroni 11 test --vip=10.0.3.2

Similar to Patroni, we start the service for each instance:

systemctl start vip-manager@11-test

The output of patronictl shows that pg1 is the leader:

+---------+--------+------------+--------+---------+----+-----------+
| Cluster | Member |    Host    |  Role  |  State  | TL | Lag in MB |
+---------+--------+------------+--------+---------+----+-----------+
| 11-test |  pg1   | 10.0.3.247 | Leader | running |  1 |           |
| 11-test |  pg2   | 10.0.3.94  |        | running |  1 |         0 |
| 11-test |  pg3   | 10.0.3.214 |        | running |  1 |         0 |
+---------+--------+------------+--------+---------+----+-----------+

In journal of ‘pg1’ it can be seen that the VIP has been configured:

Jan 19 14:53:38 pg1 vip-manager[9314]: 2020/01/19 14:53:38 IP address 10.0.3.2/24 state is false, desired true
Jan 19 14:53:38 pg1 vip-manager[9314]: 2020/01/19 14:53:38 Configuring address 10.0.3.2/24 on eth0
Jan 19 14:53:38 pg1 vip-manager[9314]: 2020/01/19 14:53:38 IP address 10.0.3.2/24 state is true, desired true

If LXC containers are used, one can also see the VIP in the output of lxc-ls -f:

NAME    STATE   AUTOSTART GROUPS IPV4                 IPV6 UNPRIVILEGED
pg1     RUNNING 0         -      10.0.3.2, 10.0.3.247 -    false
pg2     RUNNING 0         -      10.0.3.94            -    false
pg3     RUNNING 0         -      10.0.3.214           -    false

The vip-manager packages are available for Debian testing (bullseye) and unstable, as well as for the upcoming 20.04 LTS Ubuntu release (focal) in the official repositories. For Debian stable (buster), as well as for Ubuntu 19.04 and 19.10, packages are available at apt.postgresql.org maintained by credativ, along with the updated Patroni packages with vip-manager integration.

Switchover Behaviour

In case of a planned switchover, e.g. pg2 becomes the new leader:

# patronictl -c /etc/patroni/11-test.yml switchover --master pg1 --candidate pg2 --force
Current cluster topology
+---------+--------+------------+--------+---------+----+-----------+
| Cluster | Member |    Host    |  Role  |  State  | TL | Lag in MB |
+---------+--------+------------+--------+---------+----+-----------+
| 11-test |  pg1   | 10.0.3.247 | Leader | running |  1 |           |
| 11-test |  pg2   | 10.0.3.94  |        | running |  1 |         0 |
| 11-test |  pg3   | 10.0.3.214 |        | running |  1 |         0 |
+---------+--------+------------+--------+---------+----+-----------+
2020-01-19 15:35:32.52642 Successfully switched over to "pg2"
+---------+--------+------------+--------+---------+----+-----------+
| Cluster | Member |    Host    |  Role  |  State  | TL | Lag in MB |
+---------+--------+------------+--------+---------+----+-----------+
| 11-test |  pg1   | 10.0.3.247 |        | stopped |    |   unknown |
| 11-test |  pg2   | 10.0.3.94  | Leader | running |  1 |           |
| 11-test |  pg3   | 10.0.3.214 |        | running |  1 |         0 |
+---------+--------+------------+--------+---------+----+-----------+

The VIP has now been moved to the new leader:

NAME    STATE   AUTOSTART GROUPS IPV4                 IPV6 UNPRIVILEGED
pg1     RUNNING 0         -      10.0.3.247          -    false
pg2     RUNNING 0         -      10.0.3.2, 10.0.3.94 -    false
pg3     RUNNING 0         -      10.0.3.214          -    false

This can also be seen in the journals, both from the old leader:

Jan 19 15:35:31 pg1 patroni[9222]: 2020-01-19 15:35:31,634 INFO: manual failover: demoting myself
Jan 19 15:35:31 pg1 patroni[9222]: 2020-01-19 15:35:31,854 INFO: Leader key released
Jan 19 15:35:32 pg1 vip-manager[9314]: 2020/01/19 15:35:32 IP address 10.0.3.2/24 state is true, desired false
Jan 19 15:35:32 pg1 vip-manager[9314]: 2020/01/19 15:35:32 Removing address 10.0.3.2/24 on eth0
Jan 19 15:35:32 pg1 vip-manager[9314]: 2020/01/19 15:35:32 IP address 10.0.3.2/24 state is false, desired false

As well as from the new leader pg2:

Jan 19 15:35:31 pg2 patroni[9229]: 2020-01-19 15:35:31,881 INFO: promoted self to leader by acquiring session lock
Jan 19 15:35:31 pg2 vip-manager[9292]: 2020/01/19 15:35:31 IP address 10.0.3.2/24 state is false, desired true
Jan 19 15:35:31 pg2 vip-manager[9292]: 2020/01/19 15:35:31 Configuring address 10.0.3.2/24 on eth0
Jan 19 15:35:31 pg2 vip-manager[9292]: 2020/01/19 15:35:31 IP address 10.0.3.2/24 state is true, desired true
Jan 19 15:35:32 pg2 patroni[9229]: 2020-01-19 15:35:32,923 INFO: Lock owner: pg2; I am pg2

As one can see, the VIP is moved within one second.

Updated Ansible Playbook

Our Ansible-Playbook for the automated setup of a three-node cluster on Debian has also been updated and can now configure a VIP if so desired:

# ansible-playbook -i inventory -e vip=10.0.3.2 patroni.yml

Questions and Help

Do you have any questions or need help? Feel free to write to info@credativ.com.

There are two ways to authenticate yourself as a client to Icinga2. On the one hand there is the possibility to authenticate yourself by username and password. The other option is authentication using client certificates. With the automated query of the Icinga2 API, the setup of client certificates is not only safety-technically advantageous, but also in the implementation on the client side much more practical.

Unfortunately, the official Icinga2 documentation does not provide a description of the exact certificate creation process. Therefore here is a short manual:

After installing Icinga2 the API feature has to be activated first:

icinga2 feature enable api

The next step is to configure the Icinga2-node as master, the easiest way to do this is with the “node-wizard” program:

icinga2 node wizard

Icinga2 creates the necessary CA certificates with which the client certificates still to be created must be signed. Now the client certificate is created:

icinga2 pki new-cert --cn  --key .key --csr .csr

The parameter cn stands for the so-called common-name. This is the name used in the Icinga2 user configuration to assign the user certificate to the user. Usually the common name is the FQDN. In this scenario, however, this name is freely selectable. All other names can also be freely chosen, but it is recommended to use a name that suggests that the three files belong together.

Now the certificate has to be signed by the CA, Icinga2:

icinga2 pki sign-csr --csr .csr --cert .crt

Finally, the API user must be created in the file “api-user.conf”. This file is located in the subfolder of each Icinga2 configuration:

object ApiUser {
client_cn = 
permissions = []
}

For a detailed explanation of the user’s assignment of rights, it is worth taking a look at the documentation.

Last but not least Icinga2 has to be restarted. Then the user can access the Icinga2 API without entering a username and password, if he passes the certificates during the query.

You can read up on the services we provide for Icinga2 righthere.

This post was originally written by Bernd Borowski.

One would think that microcode updates are basically unproblematic on modern Linux distributions. This is fundamentally correct. Nevertheless, there are always edge cases in which distribution developers may have missed something.

Using the example of Ubuntu 18.04 LTS “Bionic Beaver” in connection with the XEN Hypervisor this becomes obvious when it comes to processors microcode updates.

Ubuntu delivers updated microcode packages for both AMD and Intel. However, these are apparently not applied to the processor.

XEN Microkernel

The reason for this is not to obvious. In XEN, the host system is already paravirtualized and cannot directly influence the CPU for security reasons. Accordingly, manual attempts to change the current microcode fail.

Therefore, the XEN microkernel has to take care of the microcode patching. Instructed correctyl, it will do so at boot time.

Customize command line in Grub

For the XEN kernel to patch the microcode of the CPU, it must have access to the microcode code files at boot time and on the other hand, he must also have the order to apply them. We can achieve the latter by Grub boot loader configuration. To do so, we setup a parameter in the kernel command line.

In the case of Ubuntu 18.04 LTS, the grub configuration file can be found at /etc/default/grub.

There you should find the file xen.cfg. This is of course only the case if the XEN Hypervisor package is installed. Open the config file in your editor and look for the location of the variable GRUB_CMDLINE_XEN_DEFAULT. Add the parameter ucode=scan. In the default state, the line of the xen.cfg then should look like this:

GRUB_CMDLINE_XEN_DEFAULT="ucode=scan"

Customize Initramfs

In addition to the instruction, the microkernel of the XEN hypervisor also needs access to the respective microcode files as well as the ‘Intel Microcode Tool’, if applicable.

While the microcode packages are usually already installed correctly, the the Intel tool may had to be made accessible via sudo apt-get install iucode tool. Care must also be taken to ensure that the microcode files also get into the initial ramdisk. For this purpose, Ubuntu already has matching scripts available.

In the default state, the system tries to select the applicable microcodes for the CPU in the InitramFS. Unfortunately, this does not succeed always, so you might have to help here.

With the command sudo lsinitrd /boot/initrd.img-4.15.0-46-generic you can, for example, check which contents are stored in the InitramFS with the name initrd.img-4.15.0-46-generic. If on an Intel system there is something from AMD but not Intel shown, the automatic processor detection went wrong when creating the initramdisk.

To get this right, you need to look at the files amd64-microcode and intel-microcode in the directory /etc/default. Each of these two config files has a INITRAMFS variable AMD64UCODE_INITRAMFS or IUCODE_TOOL_INITRAMFS. The valid values to configure are “no,” “auto,” or “early”. Default is “auto”. With “auto” the system tries the auto discovery mentioned above. If it doesn’t work, you should set the value to early in the file matching the manufacturer of your CPU, and the other setup file to no. If the manufacturer is Intel, you can use the file intel-microcode to set the following additional variable:

IUCODE_TOOL_SCANCPUS=yes

This causes the script set to perform advanced CPU detection based on the Intel CPU, so that only the microcode files are included in the InitramFS that match the CPU. This helps avoiding an oversized initial ramdisk.

Finalize changes

Both the changes to the grub config and the adjustments to the InitramFS must also be finalized. This is done via

sudo update-initramfs -u
sudo update-grub

A subsequent restart of the hypervisor will then let the XEN microkernel integrate the microcode patches provided in the InitramFS to the CPU.
Is it worth the effort?

Adjustments to the microcode of the processors are important. CPU manufacturers troubleshoot the “hardware” they sell. This fixes can be very important to maintain the integrity oder security of your server system – as we saw last year when the Spectre and Meltdown bugs got undisclosed. Of course, microcode updates can also be seen as negative, since the fixes for “Spectre” as well as “Meltdown” impose performance losses. Here it is necessary to consider whether one should integrate the microcode updates or not. This depends on risk vs. reward. Here there are quite different views, which are to be considered in the context of the system application.

A virtualization host, which runs third party virtual machines has whole other security requirements than a hypervisor who is deeply digged into the internal infrastructure and only runs trusted VMs. Between these two extremes, there are, of course, a few shades to deal with.

#Use the following YAMLs to create a cStor Storage Pool.
# and associated storage class.
apiVersion: openebs.io/v1alpha1
kind: StoragePoolClaim
metadata:
 name: cstor-disk
spec:
 name: cstor-disk
 type: disk
 poolSpec:
 poolType: striped
 # NOTE — Appropriate disks need to be fetched using `kubectl get disks`
 #
 # `Disk` is a custom resource supported by OpenEBS with `node-disk-manager`
 # as the disk operator
# Replace the following with actual disk CRs from your cluster `kubectl get disks`
# Uncomment the below lines after updating the actual disk names.
 disks:
 diskList:
# Replace the following with actual disk CRs from your cluster from `kubectl get disks`
# — disk-184d99015253054c48c4aa3f17d137b1
# — disk-2f6bced7ba9b2be230ca5138fd0b07f1
# — disk-806d3e77dd2e38f188fdaf9c46020bdc
# — disk-8b6fb58d0c4e0ff3ed74a5183556424d
# — disk-bad1863742ce905e67978d082a721d61
# — disk-d172a48ad8b0fb536b9984609b7ee653
 — -

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: openebs-postgres
  annotations:
    openebs.io/cas-type: cstor
    cas.openebs.io/config: |
      - name: StoragePoolClaim
        value: "cstor-disk"
      - name: ReplicaCount
        value: "3"       
provisioner: openebs.io/provisioner-iscsi
reclaimPolicy: Delete
---

git clone https://github.com/zalando/postgres-operator.git
cd postgres-operator

nano manifests/minimal-postgres-manifest.yaml

apiVersion: "acid.zalan.do/v1"
kind: postgresql
metadata:
  name: acid-minimal-cluster
  namespace: default
spec:
  teamId: "ACID"
  volume:
    size: 1Gi
    storageClass: openebs-postgres
  numberOfInstances: 2
  users:
    # database owner
    zalando:
    - superuser
    - createdb
 
# role for application foo
    foo_user: []
 
#databases: name->owner
  databases:
    foo: zalando
  postgresql:
    version: "10"
    parameters:
      shared_buffers: "32MB"
      max_connections: "10"
      log_statement: "all"

kubectl create -f manifests/configmap.yaml # configuration
kubectl create -f manifests/operator-service-account-rbac.yaml # identity and permissions
kubectl create -f manifests/postgres-operator.yaml # deployment

kubectl create namespace test
kubectl config set-context $(kubectl config current-context) — namespace=test

kubectl create -f manifests/minimal-postgres-manifest.yaml

$ kubectl get pv
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
pvc-8852ceef-48fe-11e9–9897–06b524f7f6ea 1Gi RWO Delete Bound default/pgdata-acid-minimal-cluster-0 openebs-postgres 8m44s
pvc-bfdf7ebe-48fe-11e9–9897–06b524f7f6ea 1Gi RWO Delete Bound default/pgdata-acid-minimal-cluster-1 openebs-postgres 7m14s

sudo apt-get install postgresql-client

kubectl get service — namespace default |grep acid-minimal-cluster

export PGHOST=$(kubectl get svc -n default -l application=spilo,spilo-role=master -o jsonpath="{.items[0].spec.clusterIP}")
export PGPORT=[HostPort]
export PGPASSWORD=$(kubectl get secret -n default postgres.acid-minimal-cluster.credentials -o ‘jsonpath={.data.password}’ | base64 -d)
psql -U postgres -c ‘create table foo (id int)’

Patroni is a PostgreSQL high availability solution with a focus on containers and Kubernetes. Until recently, the available Debian packages had to be configured manually and did not integrate well with the rest of the distribution. For the upcoming Debian 10 “Buster” release, the Patroni packages have been integrated into Debian’s standard PostgreSQL framework by credativ. They now allow for an easy setup of Patroni clusters on Debian or Ubuntu.

Patroni employs a “Distributed Consensus Store” (DCS) like Etcd, Consul or Zookeeper in order to reliably run a leader election and orchestrate automatic failover. It further allows for scheduled switchovers and easy cluster-wide changes to the configuration. Finally, it provides a REST interface that can be used together with HAProxy in order to build a load balancing solution. Due to these advantages Patroni has gradually replaced Pacemaker as the go-to open-source project for PostgreSQL high availability.

However, many of our customers run PostgreSQL on Debian or Ubuntu systems and so far Patroni did not integrate well into those. For example, it does not use the postgresql-common framework and its instances were not displayed in pg_lsclusters output as usual.

Integration into Debian

In a collaboration with Patroni lead developer Alexander Kukushkin from Zalando the Debian Patroni package has been integrated into the postgresql-common framework to a large extent over the last months. This was due to changes both in Patroni itself as well as additional programs in the Debian package. The current Version 1.5.5 of Patroni contains all these changes and is now available in Debian “Buster” (testing) in order to setup Patroni clusters.

The packages are also available on apt.postgresql.org and thus installable on Debian 9 “Stretch” and Ubuntu 18.04 “Bionic Beaver” LTS for any PostgreSQL version from 9.4 to 11.

The most important part of the integration is the automatic generation of a suitable Patroni configuration with the pg_createconfig_patroni command. It is run similar to pg_createcluster with the desired PostgreSQL major version and the instance name as parameters:

pg_createconfig_patroni 11 test

This invocation creates a file /etc/patroni/11-test.yml, using the DCS configuration from /etc/patroni/dcs.yml which has to be adjusted according to the local setup. The rest of the configuration is taken from the template /etc/patroni/config.yml.in which is usable in itself but can be customized by the user according to their needs. Afterwards the Patroni instance is started via systemd similar to regular PostgreSQL instances:

systemctl start patroni@11-test

A simple 3-node Patroni cluster can be created and started with the following few commands, where the nodes pg1, pg2 and pg3 are considered to be hostnames and the local file dcs.yml contains the DCS configuration:

for i in pg1 pg2 pg3; do ssh $i 'apt -y install postgresql-common'; done
for i in pg1 pg2 pg3; do ssh $i 'sed -i "s/^#create_main_cluster = true/create_main_cluster = false/" /etc/postgresql-common/createcluster.conf'; done
for i in pg1 pg2 pg3; do ssh $i 'apt -y install patroni postgresql'; done
for i in pg1 pg2 pg3; do scp ./dcs.yml $i:/etc/patroni; done
for i in pg1 pg2 pg3; do ssh @$i 'pg_createconfig_patroni 11 test' && systemctl start patroni@11-test'; done

Afterwards, you can get the state of the Patroni cluster via

ssh pg1 'patronictl -c /etc/patroni/11-patroni.yml list'
+---------+--------+------------+--------+---------+----+-----------+
| Cluster | Member |    Host    |  Role  |  State  | TL | Lag in MB |
+---------+--------+------------+--------+---------+----+-----------+
| 11-test |  pg1   | 10.0.3.111 | Leader | running |  1 |           |
| 11-test |  pg2   | 10.0.3.41  |        | stopped |    |   unknown |
| 11-test |  pg3   | 10.0.3.46  |        | stopped |    |   unknown |
+---------+--------+------------+--------+---------+----+-----------+

Leader election has happened and pg1 has become the primary. It created its instance with the Debian-specific pg_createcluster_patroni program that runs pg_createcluster in the background. Then the two other nodes clone from the leader using the pg_clonecluster_patroni program which sets up an instance using pg_createcluster and then runs pg_basebackup from the primary. After that, all nodes are up and running:

+---------+--------+------------+--------+---------+----+-----------+
| Cluster | Member |    Host    |  Role  |  State  | TL | Lag in MB |
+---------+--------+------------+--------+---------+----+-----------+
| 11-test |  pg1   | 10.0.3.111 | Leader | running |  1 |         0 |
| 11-test |  pg2   | 10.0.3.41  |        | running |  1 |         0 |
| 11-test |  pg3   | 10.0.3.46  |        | running |  1 |         0 |
+---------+--------+------------+--------+---------+----+-----------+

The well-known Debian postgresql-common commands work as well:

ssh pg1 'pg_lsclusters'
Ver Cluster Port Status Owner    Data directory                 Log file
11  test    5432 online postgres /var/lib/postgresql/11/test    /var/log/postgresql/postgresql-11-test.log

Failover Behaviour

If the primary is abruptly shutdown, its leader token will expire after a while and Patroni will eventually initiate failover and a new leader election:

+---------+--------+-----------+------+---------+----+-----------+
| Cluster | Member |    Host   | Role |  State  | TL | Lag in MB |
+---------+--------+-----------+------+---------+----+-----------+
| 11-test |  pg2   | 10.0.3.41 |      | running |  1 |         0 |
| 11-test |  pg3   | 10.0.3.46 |      | running |  1 |         0 |
+---------+--------+-----------+------+---------+----+-----------+
[...]
+---------+--------+-----------+--------+---------+----+-----------+
| Cluster | Member |    Host   |  Role  |  State  | TL | Lag in MB |
+---------+--------+-----------+--------+---------+----+-----------+
| 11-test |  pg2   | 10.0.3.41 | Leader | running |  2 |         0 |
| 11-test |  pg3   | 10.0.3.46 |        | running |  1 |         0 |
+---------+--------+-----------+--------+---------+----+-----------+
[...]
+---------+--------+-----------+--------+---------+----+-----------+
| Cluster | Member |    Host   |  Role  |  State  | TL | Lag in MB |
+---------+--------+-----------+--------+---------+----+-----------+
| 11-test |  pg2   | 10.0.3.41 | Leader | running |  2 |         0 |
| 11-test |  pg3   | 10.0.3.46 |        | running |  2 |         0 |
+---------+--------+-----------+--------+---------+----+-----------+

The old primary will rejoin the cluster as standby once it is restarted:

+---------+--------+------------+--------+---------+----+-----------+
| Cluster | Member |    Host    |  Role  |  State  | TL | Lag in MB |
+---------+--------+------------+--------+---------+----+-----------+
| 11-test |  pg1   | 10.0.3.111 |        | running |    |   unknown |
| 11-test |  pg2   | 10.0.3.41  | Leader | running |  2 |         0 |
| 11-test |  pg3   | 10.0.3.46  |        | running |  2 |         0 |
+---------+--------+------------+--------+---------+----+-----------+
[...]
+---------+--------+------------+--------+---------+----+-----------+
| Cluster | Member |    Host    |  Role  |  State  | TL | Lag in MB |
+---------+--------+------------+--------+---------+----+-----------+
| 11-test |  pg1   | 10.0.3.111 |        | running |  2 |         0 |
| 11-test |  pg2   | 10.0.3.41  | Leader | running |  2 |         0 |
| 11-test |  pg3   | 10.0.3.46  |        | running |  2 |         0 |
+---------+--------+------------+--------+---------+----+-----------+

If a clean rejoin is not possible due to additional transactions on the old timeline the old primary gets re-cloned from the current leader. In case the data is too large for a quick re-clone, pg_rewind can be used. In this case a password needs to be set for the postgres user and regular database connections (as opposed to replication connections) need to be allowed between the cluster nodes.

Creation of additional Instances

It is also possible to create further clusters with pg_createconfig_patroni, one can either assign a PostgreSQL port explicitly via the --port option, or let pg_createconfig_patroni assign the next free port as is known from pg_createcluster:

for i in pg1 pg2 pg3; do ssh $i 'pg_createconfig_patroni 11 test2 && systemctl start patroni@11-test2'; done
ssh pg1 'patronictl -c /etc/patroni/11-test2.yml list'
+----------+--------+-----------------+--------+---------+----+-----------+
| Cluster  | Member |       Host      |  Role  |  State  | TL | Lag in MB |
+----------+--------+-----------------+--------+---------+----+-----------+
| 11-test2 |  pg1   | 10.0.3.111:5433 | Leader | running |  1 |         0 |
| 11-test2 |  pg2   |  10.0.3.41:5433 |        | running |  1 |         0 |
| 11-test2 |  pg3   |  10.0.3.46:5433 |        | running |  1 |         0 |
+----------+--------+-----------------+--------+---------+----+-----------+

Ansible Playbook

In order to easily deploy a 3-node Patroni cluster we have created an Ansible playbook on Github. It automates the installation and configuration of PostgreSQL and Patroni on the three nodes, as well as the DCS server on a fourth node.

Questions and Help

Do you have any questions or need help? Feel free to write to info@credativ.com.

# add debmon
cat <<EOF >/etc/apt/sources.list.d/debmon.list
deb http://debmon.org/debmon debmon-jessie main
EOF

# add debmon key
wget -O - http://debmon.org/debmon/repo.key 2>/dev/null | apt-key add -

# update repos
apt-get update

apt-get install icinga2

# install packages for icinga2 and graphite-web and carbon

apt-get install icinga2 graphite-web graphite-carbon libapache2-mod-wsgi apache2

# add repo (package for wheezy works on jessie)
cat </etc/apt/sources.list.d/grafana.list
deb https://packagecloud.io/grafana/stable/debian/ wheezy main
EOF
 
# add key
curl -s https://packagecloud.io/gpg.key | sudo apt-key add -
 
# update repos
apt-get update

d-i debian-installer/locale string de_DE.UTF-8

d-i debian-installer/locale string de_DE.UTF-8
d-i debian-installer/keymap select de-latin1
d-i console-keymaps-at/keymap select de
d-i languagechooser/language-name-fb select German
d-i countrychooser/country-name select Germany
d-i console-setup/layoutcode string de_DE
d-i clock-setup/utc boolean true
d-i time/zone string Europe/Berlin
d-i clock-setup/ntp boolean true
d-i clock-setup/ntp-server string ntp1
 
tasksel tasksel/first multiselect standard, desktop, gnome-desktop, laptop

d-i pkgsel/include string openssh-client vim less rsync

label i386
kernel debian-installer/i386/linux
append vga=normal initrd=debian-installer/i386/initrd.gz netcfg/choose_interface=eth0 domain=example.com locale=de_DE debian-installer/country=DE debian-installer/language=de debian-installer/keymap=de-latin1-nodeadkeys console-keymaps-at/keymap=de-latin1-nodeadkeys auto-install/enable=false preseed/url=http://$server/preseed.cfg DEBCONF_DEBUG=5 -- quiet

d-i debian-installer/locale string de_DE.UTF-8
d-i debian-installer/keymap select de-latin1
d-i console-keymaps-at/keymap select de
d-i languagechooser/language-name-fb select German
d-i countrychooser/country-name select Germany
d-i console-setup/layoutcode string de_DE

# Network
d-i netcfg/choose_interface select auto
d-i netcfg/get_hostname string debian
d-i netcfg/get_domain string example.com
 
# Package mirror
d-i mirror/protocol string http
d-i mirror/country string manual
d-i mirror/http/hostname string debian.example.com
d-i mirror/http/directory string /debian
d-i mirror/http/proxy string
d-i mirror/suite string lenny
 
# Timezone
d-i clock-setup/utc boolean true
d-i time/zone string Europe/Berlin
d-i clock-setup/ntp boolean true
d-i clock-setup/ntp-server string ntp.example.com
 
# Root-Account
d-i passwd/make-user boolean false
d-i passwd/root-password password secretpassword
d-i passwd/root-password-again password secretpassword
 
# Further APT-Options
d-i apt-setup/non-free boolean false
d-i apt-setup/contrib boolean false
d-i apt-setup/security-updates boolean true
 
d-i apt-setup/local0/source boolean false
d-i apt-setup/local1/source boolean false
d-i apt-setup/local2/source boolean false
 
# Tasks
tasksel tasksel/first multiselect standard, desktop
d-i pkgsel/include string openssh-client vim less rsync
d-i pkgsel/upgrade select safe-upgrade
 
# Popularity-Contest
popularity-contest popularity-contest/participate boolean true

# Command to be followed after the installation. `in-target` means that the following
# Command is followed in the installed environment, rather than in the installation environment.
# Here http://$server/skript.sh nach /tmp is downloaded, enabled and implemented.
d-i preseed/late_command string in-target wget -P /tmp/ http://$server/skript.sh; in-target chmod +x /tmp/skript.sh; in-target /tmp/skript.sh>

[modules]
allow 192.168.0.1/24
allow *.credativ.de

mkdir -p /etc/puppet/modules/ssh/manifests
mkdir -p /etc/puppet/modules/ssh/files

class ssh{
        package { "openssh-server":
                 ensure => latest,
        }
        file { "/etc/ssh/sshd_config":
                owner   => root,
                group   => root,
                mode    => 644,
                source  => "puppet:///ssh/sshd_config",
        }
        service { ssh:
                ensure          => running,
                hasrestart      => true,
                subscribe       => File["/etc/ssh/sshd_config"],
        }
        file { "/home/credativ/.ssh":
                path    => "/home/credativ/.ssh",
                owner   => "credativ",
                group   => "credativ",
                mode    => 600,
                recurse => true,
                source  => "puppet:///ssh/ssh",
                ensure  => [directory, present],
        }
}

node default {
   include rsyslog
}
 
node 'external' inherits default {
  include ssh
}