From b9e13eea30fa89d3de8983e9cd2205ace75f2a61 Mon Sep 17 00:00:00 2001
From: Julien Dessaux
Date: Wed, 1 Sep 2021 14:53:05 +0200
Subject: Added kubernetes evicted pods article

---
 content/blog/kubernetes/delete-all-evicted-pods.md | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
 create mode 100644 content/blog/kubernetes/delete-all-evicted-pods.md

(limited to 'content/blog/kubernetes/delete-all-evicted-pods.md')

diff --git a/content/blog/kubernetes/delete-all-evicted-pods.md b/content/blog/kubernetes/delete-all-evicted-pods.md
new file mode 100644
index 0000000..ec624fd
--- /dev/null
+++ b/content/blog/kubernetes/delete-all-evicted-pods.md
@@ -0,0 +1,22 @@
+---
+title: How to delete all evicted pods in kubernetes
+description: A quick note for future reference
+date: 2021-09-01
+---
+
+## Introduction
+
+I was playing with the percona xtradb operator on one of my test clusters last week and left it in a state where mysqld errors logs were piling up over the week-end. On monday morning my nodes had their filesystems full and I discovered what kubernetes evicted pods were : pods that fail when a node's resources get constrained.
+
+My problem is : these evicted pods lingered, so I looked for a way to clean them up.
+
+## How to delete all evicted pods
+
+My google fu directed my towards several commands looking like the following, but they all had a thing or another that did not work. Here is the one I pieced together from these various resources :
+```sh
+kubectl get pods --all-namespaces -o json |
+    jq '.items[] | select(.status.reason!=null) |
+        select(.status.reason | contains("Evicted")) |
+        "kubectl -n \(.metadata.namespace) delete pod \(.metadata.name)"' |
+    xargs -n 1 bash -c
+```
-- 
cgit v1.2.3