Linux中國
使用 KRAWL 掃描 Kubernetes 錯誤
當你使用 Kubernetes 運行容器時,你通常會發現它們堆積在一起。這是設計使然。它是容器的優點之一:每當需要新的容器時,它們啟動成本都很低。你可以使用前端工具(如 OpenShift 或 OKD)來管理 Pod 和容器。這些工具使可視化設置變得容易,並且它具有一組豐富的用於快速交互的命令。
如果管理容器的平台不符合你的要求,你也可以僅使用 Kubernetes 工具鏈獲取這些信息,但這需要大量命令才能全面了解複雜環境。出於這個原因,我編寫了 KRAWL,這是一個簡單的腳本,可用於掃描 Kubernetes 集群命名空間下的 Pod 和容器,並在發現任何事件時,顯示事件的輸出。它也可用作為 Kubernetes 插件使用。這是獲取大量有用信息的快速簡便方法。
先決條件
- 必須安裝
kubectl
。 - 集群的 kubeconfig 配置必須在它的默認位置(
$HOME/.kube/config
)或已被導出到環境變數(KUBECONFIG=/path/to/kubeconfig
)。
使用
$ ./krawl
![KRAWL script](/data/attachment/album/202002/27/102959imepam5p3asfgggs.png "KRAWL script")
腳本
#!/bin/bash
# AUTHOR: Abhishek Tamrakar
# EMAIL: abhishek.tamrakar08@gmail.com
# LICENSE: Copyright (C) 2018 Abhishek Tamrakar
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
##
#define the variables
KUBE_LOC=~/.kube/config
#define variables
KUBECTL=$(which kubectl)
GET=$(which egrep)
AWK=$(which awk)
red=$(tput setaf 1)
normal=$(tput sgr0)
# define functions
# wrapper for printing info messages
info()
{
printf 'ne[34m%se[m: %sn' "INFO" "$@"
}
# cleanup when all done
cleanup()
{
rm -f results.csv
}
# just check if the command we are about to call is available
checkcmd()
{
#check if command exists
local cmd=$1
if [ -z "${!cmd}" ]
then
printf 'ne[31m%se[m: %sn' "ERROR" "check if $1 is installed !!!"
exit 1
fi
}
get_namespaces()
{
#get namespaces
namespaces=(
$($KUBECTL get namespaces --ignore-not-found=true |
$AWK '/Active/ {print $1}'
ORS=" ")
)
#exit if namespaces are not found
if [ ${#namespaces[@]} -eq 0 ]
then
printf 'ne[31m%se[m: %sn' "ERROR" "No namespaces found!!"
exit 1
fi
}
#get events for pods in errored state
get_pod_events()
{
printf 'n'
if [ ${#ERRORED[@]} -ne 0 ]
then
info "${#ERRORED[@]} errored pods found."
for CULPRIT in ${ERRORED[@]}
do
info "POD: $CULPRIT"
info
$KUBECTL get events
--field-selector=involvedObject.name=$CULPRIT
-ocustom-columns=LASTSEEN:.lastTimestamp,REASON:.reason,MESSAGE:.message
--all-namespaces
--ignore-not-found=true
done
else
info "0 pods with errored events found."
fi
}
#define the logic
get_pod_errors()
{
printf "%s %s %sn" "NAMESPACE,POD_NAME,CONTAINER_NAME,ERRORS" > results.csv
printf "%s %s %sn" "---------,--------,--------------,------" >> results.csv
for NAMESPACE in ${namespaces[@]}
do
while IFS=' ' read -r POD CONTAINERS
do
for CONTAINER in ${CONTAINERS//,/ }
do
COUNT=$($KUBECTL logs --since=1h --tail=20 $POD -c $CONTAINER -n $NAMESPACE 2>/dev/null|
$GET -c '^error|Error|ERROR|Warn|WARN')
if [ $COUNT -gt 0 ]
then
STATE=("${STATE[@]}" "$NAMESPACE,$POD,$CONTAINER,$COUNT")
else
#catch pods in errored state
ERRORED=($($KUBECTL get pods -n $NAMESPACE --no-headers=true |
awk '!/Running/ {print $1}' ORS=" ")
)
fi
done
done< <($KUBECTL get pods -n $NAMESPACE --ignore-not-found=true -o=custom-columns=NAME:.metadata.name,CONTAINERS:.spec.containers[*].name --no-headers=true)
done
printf "%sn" ${STATE[@]:-None} >> results.csv
STATE=()
}
#define usage for seprate run
usage()
{
cat << EOF
USAGE: "${0##*/} </path/to/kube-config>(optional)"
This program is a free software under the terms of Apache 2.0 License.
COPYRIGHT (C) 2018 Abhishek Tamrakar
EOF
exit 0
}
#check if basic commands are found
trap cleanup EXIT
checkcmd KUBECTL
#
#set the ground
if [ $# -lt 1 ]; then
if [ ! -e ${KUBE_LOC} -a ! -s ${KUBE_LOC} ]
then
info "A readable kube config location is required!!"
usage
fi
elif [ $# -eq 1 ]
then
export KUBECONFIG=$1
elif [ $# -gt 1 ]
then
usage
fi
#play
get_namespaces
get_pod_errors
printf 'n%40sn' 'KRAWL'
printf '%sn' '---------------------------------------------------------------------------------'
printf '%sn' ' Krawl is a command line utility to scan pods and prints name of errored pods '
printf '%snn' ' +and containers within. To use it as kubernetes plugin, please check their page '
printf '%sn' '================================================================================='
cat results.csv | sed 's/,/,|/g'| column -s ',' -t
get_pod_events
此文最初發布在 KRAWL 的 GitHub 倉庫下的 README 中,並被或許重用。
via: https://opensource.com/article/20/2/kubernetes-scanner
作者:Abhishek Tamrakar 選題:lujun9972 譯者:geekpi 校對:wxy
本文轉載來自 Linux 中國: https://github.com/Linux-CN/archive
對這篇文章感覺如何?
太棒了
0
不錯
0
愛死了
0
不太好
0
感覺很糟
0
More in:Linux中國
如何通過 VLC 使用字幕
使用 VLC 媒體播放器播放和管理字幕的新手指南。
Unix 桌面:在 Linux 問世之前
僅僅開源還不足以實現開放,還需開放標準和建立共識。
Valve 對於 Ubuntu 的 Snap 版本的 Steam 並不滿意:原因何在
你可能會發現,Snap 版本的 Steam 並不如你期待的那樣好,你怎麼看?
Wine 9.0 發布,實驗性地加入了 Wayland 驅動
Wine 的這個新版本正在為未來做好準備!