Posts

When deploying the prometheus-community/kube-prometheus-stack in a k8s cluster, the default retention for Prometheus scrapped data is 10d at the time of this writing or Helm chart kube-prometheus-stack-69.4.1. As a result, the longhorn storage is filled up fast, and caused the Prometheus to crash. So I’d decided to change the retention period to 3 days, and also set retention size to 4GB. To do this, update the values file for the Helm chart with the following values: ...

This is a quick note on how to reset the password of user account on the open sourced Xen Orchestra. To reset the password of a user, of course, you’d need to know the user account, e.g. admin@admin.net. [root@xoa1 xo-server]# cd /opt/xo/xo-server [root@xoa1 xo-server]# ./dist/recover-account-cli.mjs --help xo-server-recover-account <user name or email> If the user does not exist, it is created, if it exists, updates its password, remove any configured OTP and resets its permission to Admin. [root@xoa1 xo-server]# ./dist/recover-account-cli.mjs admin@admin.net Password (leave empty for random): ******* user admin@admin.net has been successfully updated However, if you also forgot the account itself, you can create a new user with the same command. Here is an example to create a new account called admin: ...

Up until recently, my preferred method of creating Podman containers is to use Ansible, and that has been working fine for me. However, I just stumble the new-to-me method called “quadlet”. So far, I really like it, and already started moving some containers to quadlet. To learn more about Quadlet, check out [1]. But, today I’m going to share an example of creating a simple container with Quadlet. In this example, the container whoami is run in rootless mode, under a non-privileged user. ...

If you use ZFS as the file system, you probably already know that tank is the most common name for ZFS pools. That also applies to me, I do name most of my ZFS pools as tank as the default. There is no issue with that name, however, since ZFS’s pool name must be unique, this means I couldn’t import or have multiple ZFS pools with the name tank on a system. ...

“The systemd login manager will terminate any user services when their login session ends unless you enable user lingering with the enable-linger option. Use loginctl with the enable-linger option to configure a user to start during the boot process and prevent the systemd service from terminating when the user session ends.” [1] To enable linger for a user, e.g. kenno, we can run the following command: # loginctl enable-linger kenno I usually ran the above command, and that was it. However, today I need to check whether the user kenno had linger enabled already. I also want to document this for my own future reference. ...