<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>High-Availability on 5glinux</title><link>https://blog.5glinux.com/tags/high-availability/</link><description>Recent content in High-Availability on 5glinux</description><image><title>5glinux</title><url>https://blog.5glinux.com/images/og-default.jpg</url><link>https://blog.5glinux.com/images/og-default.jpg</link></image><generator>Hugo</generator><language>en-us</language><lastBuildDate>Tue, 03 Nov 2026 20:00:00 +0000</lastBuildDate><atom:link href="https://blog.5glinux.com/tags/high-availability/index.xml" rel="self" type="application/rss+xml"/><item><title>Stretched SUSE Virtualization cluster: what 50 ms to the witness site does to etcd</title><link>https://blog.5glinux.com/posts/stretched-suse-virtualization-cluster-witness-latency/</link><pubDate>Tue, 03 Nov 2026 20:00:00 +0000</pubDate><guid>https://blog.5glinux.com/posts/stretched-suse-virtualization-cluster-witness-latency/</guid><description>A stretched SUSE Virtualization cluster needs a third site for etcd quorum. I put 25 ms each way between that site and the other two nodes with tc netem and measured the control plane in four scenarios: normal operation, witness as leader, a data center lost, and the worst case with a VM create.</description></item></channel></rss>