User Tools

Site Tools


server_administration

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
server_administration [2016/05/26 21:07]
sgripon [fail2ban and UFW]
server_administration [2016/12/28 12:06] (current)
sgripon [logwatch]
Line 1: Line 1:
-====== Server Administration ======+====== ​Linux Server Administration ======
  
 ~~socialite~~ ~~socialite~~
Line 67: Line 67:
 </​file>​ </​file>​
  
-===== ddos-deflate =====+Then activate ssh jail by modifying ///​etc/​fail2ban/​jail.conf//​. I also decided to ban for 1 year attackers:
  
-This is a simple script that automatically ban ip when the number of connections exceed what is configuredSee https://​antiddos.eu/​en/​news/​item/​20.+<code ini> 
 +...
  
-**Note**: if you have an error "$CONF not found" when running the script, you mau change the first line of the script from:+# Ignore my own IP in order to avoid being locked outside 
 +ignoreip = 127.0.0.1/8 xxx.xxx.xxx.xxx
  
-  ​#!/bin/sh +"​bantime"​ is the number of seconds that a host is banned. 
-   +# 1 year 
-to:+bantime ​ = 31536000
  
-  ​#!/bin/bash +Default banning action 
-   +banaction = ufw-ssh
-Same thing in the cron job.+
  
-There is also an issue with cron service nameRaplace in the file all occurrences of  +# Activate ssh jail 
-  service ​crond restart+[ssh] 
 +enabled ​ = true 
 +port     = ssh 
 +filter ​  = sshd 
 +logpath ​ = /​var/​log/​auth.log 
 +maxretry = 6 
 + 
 +[ssh-ddos] 
 +enabled ​ = true 
 +port     = ssh 
 +filter ​  = sshd-ddos 
 +logpath ​ = /​var/​log/​auth.log 
 +maxretry = 6 
 + 
 +... 
 +</​code>​ 
 + 
 +Finally, reload configuration:​ 
 + 
 +  sudo service ​fail2ban ​restart 
 + 
 +===== Under DDOS attack? ===== 
 + 
 +The following command can help you identify ddos attacks and IP adresses which are at the source of the attack: 
 + 
 +  netstat -ntu | awk '​{if(NR>​2)print $5}' | cut -d: -f1 | sort | uniq -c | sort -n
   ​   ​
-with +It will print for each connected IP the count of open connections:​ 
-  ​service ​cron restart+ 
 +      5 xxx.xxx.xxx.xxx 
 +      158 yyy.yyy.yyy.yyy 
 +       
 +Usually when you have a high number of open connections,​ like here for yyy.yyy.yyy.yyy,​ it is probably that this ip tries to DDOS you. That's time to ban it using ufw. 
 +===== nmd ===== 
 + 
 +No More DDOS (nmd) is a simple script that automatically ban ip when the number of connections exceed what is configured. See http://​us.informatiweb-pro.net/​system-admin/​linux/​17--debian-ubuntu-centos-block-ddos-attacks-with-no-more-ddos-formerly-ddos-deflate.html. (By Lionel Eppe) 
 + 
 +I modified a little the script in order to use ufw to ban adresses. Also, there is an issue with the installed cron script: 
 +  ​- The name of the cron script must not contain dot (modify CRON variable in ///​usr/​local/​nmd/​ndm.conf/​agent.conf//​. 
 +  - The second cron command in the file miss the user root 
 + 
 +The good cron script must look like this: 
 + 
 +<file bash /​etc/​cron.d/​nmd>​ 
 +* * * * * root /​usr/​local/​nmd/​nmd-agent.sh >> /​var/​log/​nmd-agent.log 2>&​1 
 +0 0 */7 * 0 root echo  > /​var/​log/​nmd-agent.log 2>&​1 
 +</​file>​
  
 ==== Use IPTables to limit NEW traffic on port 80 and 443 ==== ==== Use IPTables to limit NEW traffic on port 80 and 443 ====
Line 133: Line 177:
   ​   ​
 After that, you should be able to login via ssh without the need of the password. After that, you should be able to login via ssh without the need of the password.
 +
 +===== logwatch =====
 +
 +Logwatch can send you a formatted view of system logs every morning by email. It is usefull to check everyday the health of your server.
 +
 +First, copy default config file to change settings:
 +
 +  sudo cp /​usr/​share/​logwatch/​default.conf/​logwatch.conf /​etc/​logwatch/​conf/​
 +
 +Then, to receive by email the report modify the cron job and add --mailto option :
 +
 +<file bash /​etc/​cron.daily/​00logwatch>​
 +#!/bin/bash
 +
 +#Check if removed-but-not-purged
 +test -x /​usr/​share/​logwatch/​scripts/​logwatch.pl || exit 0
 +
 +#execute
 +/​usr/​sbin/​logwatch --output mail --mailto my.email@domain.tld
 +
 +</​file>​
 +
 +===== Slow server diagnostic =====
 +
 +See a very good flow chart to help in slow server cases : http://​blog.scoutapp.com/​articles/​2014/​07/​31/​slow_server_flow_chart
  
 **Share this page:** **Share this page:**
server_administration.1464289656.txt.gz · Last modified: 2016/05/26 21:07 by sgripon