Web Hosting Journal
Sometimes, you may want to run a screen command in a remote server. That makes it necessary to run the command inside the screen session while starting it.
# screen -d -m sh -c "yourcommand"
From the man page of Screen:
-d -m : Start screen in “detached” mode. This creates a new session but doesn’t attach to it. This is useful for system startup scripts.
sh -c: Starts a shell and runs a command for you.
How about install Let’s Encrypt for Cpanel?
Before we start, you may first want to install Let’s Encrypt to use an Immediately Issuing provider for your SSL in Cpanel?
https://mellowhost.com/blog/how-to-install-lets-encrypt-in-cpanel.html
Once done, you may now continue using this tutorial to install Let’s Encrypt for your Service SSL in Cpanel/WHM/Webmail.
How To Install AutoSSL for Server Hostname / Webmail / Cpanel / WHM
Starting from Cpanel 11.58, Cpanel is offering Free SSL, issued by ‘Cpanel INC’ for free of charge to the valid cpanel license owner. If you are using cpanel, login to your WHM >> Providers >> Enable Cpanel & from Options >> Check Allow AutoSSL to replace invalid or expiring non-AutoSSL certificates.
Now, running upcp should automatically install the free SSL for your cpanel server hostname. If it doesn’t, it is probably because your server IP and the hostname IP are resolving to wrong address. To understand and troubleshoot the problem, run the following script from command line:
# /usr/local/cpanel/bin/checkallsslcerts
This script checks and installs certificate for expired, invalid and self signed certificates for the server services. If you are seeing an error like the following:
[WARN] The system failed to acquire a signed certificate from the cPanel Store because of the following error: (XID 62hp6x) The system queried for a temporary file at “http://server91.mellowhost.com/.well-known/pki-validation/D92868E512FB02354F2498B94E67430B.txt”, but the web server responded with the following error: 404 (Not Found). A DNS (Domain Name System) or web server misconfiguration may exist.
It means, your hostname is resolving to the wrong IP. You would need to check if the hostname is resolving to an IP which has first virtualhost pointed to /var/www/html or not under /etc/apache2/conf/httpd.conf
You must upgrade to Litespeed “5.2.1 build 2 or later”, in order to upgrade to the next version of cPanel & WHM.
The error appears because Litespeed below 5.2.1 doesn’t have SSL compatibility with Cpanel 11.68. Although, this goes further, if you uninstall the plugin, and upcp will still fail to update the Cpanel/WHM. A workaround for this, is to install and uninstall the plugin through Litespeed auto installer. Here is how to do that:
# wget https://www.litespeedtech.com/packages/cpanel/lsws_whm_autoinstaller.sh
# chmod a+x lsws_whm_autoinstaller.sh
# ./lsws_whm_autoinstaller.sh TRIAL 1 8080 username testpass1234 [email protected] 1 0
# /usr/local/lsws/admin/misc/uninstall.sh
# /usr/local/cpanel/whostmgr/docroot/cgi/lsws/lsws_whm_plugin_uninstall.sh
# /scripts/upcp --force
I had a client ticket today, with the following screenshot:
The error says, ‘server IP address could not be found’. This type of error means there is a DNS resolution error. There could be 3 possibilities:
To check if the client has updated the proper dns nameservers, you can use intodns.com. It will also tell you if the host DNS is down or not. If both are ok, you should check if you are able to load other domains using your Internet, if not, it has things to do with the local DNS resolver of your desktop or the ISP. In my case, it was client who failed to update the nameservers of the domain. All that was required to update the nameservers with the server ones.
Exim provides a quick way to check the total number of mails in the queue. This is done using the exim -bpc Although, this is not the same for postfix. Postfix doesn’t come with an easy way to do that.
A quick tip on what I use to check the postfix queue number is the following command:
# mailq | tail -n 1
-- 6899 Kbytes in 1518 Requests.
Basically, postfix returns the queue statistics at the end of the queue listing command. We are simply tailing that to find the number.
Postfix queue can be cleared using the postsuper command. The syntax is as following:
# postsuper -d ALL
It will clear all the postfix queue. You can specifically clear the deferred emails for example from the postfix queue as following:
# postsuper -d ALL deferred
If you have a large quantity in postfix queue, there is actually a quicker and easier way to do that. You can simply remove all the folders under /var/spool/postfix and it will clear the queue.
# rm -Rf /var/spool/postfix/*
# mailq | tail -n 1
Mail queue is empty
Note: The folders inside postfix would get created automatically once the queue starts filling up, nothing to worry about.
Exim queue can be checked using the following:
# exim -bp
To check the number of mails in queue, you can use:
# exim -bpc
To remove a message from exim queue, you need to use the following:
# exim -Mrm {message-id}
There is no build in command to clear all the mails from exim queue. You can use a pipe command to clear the exim queue as following:
# exim -bp | exiqgrep -i | xargs exim -Mrm
Although, there are even quicker and easier way to clear the exim queue, specially if you have a lot of emails in queue and the server is pretty loaded.
# rm -Rf /var/spool/exim/input
Removing the input directory should clear the exim queue. Note: The directory would automatically create once the exim starts it’s queue again, no need to worry.
One of my customer came with an error saying the postfix in his server isn’t working. The server was running CentOS 7, and the system postfix status was inactive, means not running. Although, the system queue was running I could see. The error that was returning while restarting/checking status was the following:
# service postfix status
Redirecting to /bin/systemctl status postfix.service
● postfix.service - Postfix Mail Transport Agent
Loaded: loaded (/usr/lib/systemd/system/postfix.service; enabled; vendor preset: disabled)
Active: failed (Result: exit-code) since Tue 2018-01-09 04:04:05 UTC; 1s ago
Process: 9201 ExecStart=/usr/sbin/postfix start (code=exited, status=1/FAILURE)
Process: 9197 ExecStartPre=/usr/libexec/postfix/chroot-update (code=exited, status=0/SUCCESS)
Process: 9194 ExecStartPre=/usr/libexec/postfix/aliasesdb (code=exited, status=0/SUCCESS)
Main PID: 1358 (code=killed, signal=TERM)
Jan 09 04:04:03 twin7.hifrank.biz systemd[1]: Starting Postfix Mail Transport Agent...
Jan 09 04:04:03 twin7.hifrank.biz postfix/master[9273]: fatal: open lock file /var/lib/postfix/master.lock: unable to set exclusive lock: Resource tempo...vailable
Jan 09 04:04:04 twin7.hifrank.biz postfix/master[9272]: fatal: daemon initialization failure
Jan 09 04:04:05 twin7.hifrank.biz postfix/postfix-script[9274]: fatal: mail system startup failed
Jan 09 04:04:05 twin7.hifrank.biz systemd[1]: postfix.service: control process exited, code=exited status=1
Jan 09 04:04:05 twin7.hifrank.biz systemd[1]: Failed to start Postfix Mail Transport Agent.
Jan 09 04:04:05 twin7.hifrank.biz systemd[1]: Unit postfix.service entered failed state.
Jan 09 04:04:05 twin7.hifrank.biz systemd[1]: postfix.service failed.
The error to note here is the following:
fatal: open lock file /var/lib/postfix/master.lock
I first killed the smtp and smtpd processes that runs by postfix:
# killall -9 smtp
# killall -9 smtpd
But that didn’t solve the problem. I then used the fuser command to check which process holds the lock file:
# fuser /var/lib/postfix/master.lock
/var/lib/postfix/master.lock: 18698
Then we check the process 18698 and kill the responsible process:
# ps -axwww|grep 18698
9333 pts/0 S+ 0:00 grep --color=auto 18698
18698 ? Ss 4:28 /usr/libexec/postfix/master -w
# killall -9 /usr/libexec/postfix/master
or
# kill -9 18698
Once the process is killed, you can now start the postfix:
# service postfix start
# service postfix status|grep Active
Redirecting to /bin/systemctl status postfix.service
Active: active (running) since Tue 2018-01-09 04:15:50 UTC; 4min 45s ago
If you are using a centos 7 minimal installation, and trying to kill process by name using the command ‘killall’, you are most likely going to see the error:
# killall -9 php
killall: command not found
The error appears because CentOS 7 encourages you to use pkill instead of killall to kill process by name. pkill has versatile application, although, it can be used to kill process by name same as killall.
You can use pkill. pkill is a simple command. It’s syntax is as following:
# pkill processname
For example, if you want to kill all the php process, run:
# pkill php
Note: It will kill all the processes that match php. To list the process that pkill going to kill, you can use pgrep as following:
# pgrep -l php
If you do not want to use pkill, and keep using killall commands in centos 7, this is also possible. killall is a part of psmisc yum package. All you have to do, is to install psmisc in your system using yum
# yum install psmisc
# killall -9 php
CentOS 7 doesn’t come with ifconfig tools. It encourages users to use ‘ip’ tool for network administration. Although, it is still possible to use ifconfig with CentOS 7. ifconfig is a part of net-tools package. All you have to do is to install the net-tools package using yum.
Run the following command to install net-tools package in CentOS 7, this will install ifconfig as well:
# yum install net-tools -y
# ifconfig
In CentOS 6, timezone files are located under /usr/share/zoneinfo. So, if your zone is for example, America/Chicago (UTC -6), it would be /usr/share/zoneinfo/America/Chicago and so on.
CentOS 6, uses a file called ‘localtime’ located under /etc to determine it’s currently set timezone.
# ls -la /etc/localtime
This file, is either the actual time zone file moved to this location or a symlink to the timezone under zoneinfo directory. So if you want to change the timezone, first you need to determine which timezone to use and then symlink it to localtime. You can do that using the following:
# rm -f /etc/localtime
# ln -s /usr/share/zoneinfo/Asia/Dhaka /etc/localtime
# date
This would set the current timezone to GMT +6 BDT or Asia/Dhaka timezone, zone I belong to.
CentOS 7, comes with a tool called ‘timedatectl’. This can be used to find and set the symlink for you instead of doing the work that were required in CentOS 6.
To list available time zones, run:
# timedatectl list-timezones
You can find your desired timezone, as following:
# timedatectl list-timezones | grep Chicago
Now, to set a time zone, use the command set-timezone with timedatectl command. For example, if I want to set the time zone to America/Chicago, I would run the following:
# timedatectl set-timezone America/Chicago
# date
This also should create the symlink of locatime file to the zoneinfo directory. You can see that with the following:
# ls -l /etc/localtime