Category: Sysadmin

URL re-writing with nginx, PHP, and WordPress

There are many posts about nginx, re-directs, PHP, and WordPress. There are somewhat fewer posts that talk about (internal) re-writes, where the request by the web browser is mangled to be served by another resource than the one requested.

For example, I may want a request for https://mysite.foo/cool/penguin to actually be served by https://mysite.foo/coolstuff.php?id=penguin, or simply setup an alias such as https://mysite.foo/cool/penguin to be served by https://mysite.foo/cool/linux, but preserve the URL in the browser address bar.

With PHP-FPM and nginx, you run into an additional problem, which is the fastcgi_parm variables that are passed from nginx to PHP-FPM. So even if you have really fancy URL re-writing configured (and working), the end result may not be passed on to PHP-FPM from nginx.

So solve this, you should look into this construct, which is present in many nginx configurations as a default setup:

fastcgi_param REQUEST_URI $request_uri;

Since your needs probably differ from mine, I wont make this post any longer than it has to be, but that fastcgi_param line above may be a good starting point if you’re experiencing problems with nginx, PHP-FPM, and URL re-writing.

Good luck!

What’s My IP?

There are a number of ways to figure out your public IP address automatically, which can be extremely useful for Dynamic DNS (DDNS) situations or other automation ventures, these are some of them:

dig +short @resolver1.opendns.com ANY myip.opendns.com
curl https://ifconfig.co
curl http://whatismyip.akamai.com/

It would be nice if the Cloudflare DNS service (1.1.1.1) supported this too!

Changing your MTU may help certbot / Let’s Encrypt

While attempting to create a Let’s Encrypt SSL certificate using certbot for a number of domains, I ran into something that appeared to be a timeout issue in the certbot client and/or one of the Python libraries used by it.

I found this thread, which recommends changing the MTU to 1300, and it does work. So, maybe this will help you too 🙂

community.letsencrypt.org/t/cannot-get-new-certificate-readtimeout-error/94586

In other words:

ifconfig eth0 mtu 1300

may help you out.

Things to do after migrating to Zimbra 8.8.x

As I recently handled a migration from Zimbra 8.7.11 to Zimbra 8.8.12, I’ve been running into things I think the Zimbra updater should handle by itself, and/or things that should be documented in a more practical way than they currently are. I’ll just keep updating this post with things I find and learn.

(The migration turned into an 18 hour job since we ran into a +1 year old problem where an Ubuntu upgrade form 14.04.LTS to 16.04.LTS more or less wipes your Zimbra setup and leaves it in a very sorry state. Fortunately, I managed to piece it back by hand.)

Zimbra Backup NG

The new and improved Zimbra Backup NG should be enabled IMHO. It does things a lot better than the previous version(s). In a number of places, Zimbra makes references to its improvements, and so on. They should probably include this instruction too:

To disable the “classic” (or “legacy”) backups, a simple

zmschedulebackup -F

(as the “zimbra” user) is all you need to execute. (You may want to hang on to your old backups until they’ve outlived their purpose though. This command only disables/removes the scheduled “legacy” backups.)

Using sfdisk to recover a partition table on Linux

As he re-entered the sfdisk dump manually in the remote recovery console, using the devil’s editor (vi), he was silently thanking the Linux developers for not screwing around with the file system when it cannot be mounted.

Messing around with partition tables, disk volumes, and similar critical configuration parameters can lead to quite unexpected and unintended results. So, it may be a good idea to actually dump the current configuration before you begin your magic.

Using sfdisk, you can dump your Linux partition configuration in a fairly straightforward way. You can try the command by typing just sfdisk -d /dev/disk, where disk is one of the disks in your Linux system. For a list of disks in your system, use the lsblk command. They are identified as “disk” (surprise).

sfdisk -d /dev/sda > sda.txt

This would dump the partition table data for the /dev/sda disk to the file sda.txt. Your output will look something like this:

label: dos
label-id: 0xa828a5d8
device: /dev/sda
unit: sectors

/dev/sda1 : start= 2048, size= 997376, type=83, bootable
/dev/sda2 : start= 999424, size= 999424, type=82
/dev/sda3 : start= 1998848, size= 249659359, type=83

The partition table information can then later be restore by issuing the reverse, i.e.

sfdisk /dev/sda < sda.txt

DO NOT PERFORM THE ABOVE COMMAND IF YOU DON’T KNOW WHAT YOU ARE DOING!

This procedure may come in handy if you, like me, manage to screw up the partition table and find yourself at the (initramfs) prompt when you restart your Linux machine. You will (obviously) need to save the dump file (sda.txt above) in a location other than your computer. Using this method, it’s often possible to recover your partition table from a rescue boot (be it on CD, DVD or a flash drive).

I happened to have a previous terminal session window open with the above information, so I hand-typed it from one window to another, where I was running the remote recovery console.

There are a lot more complex partition setups than the above, and sfdisk may not work in those cases or for certain RAID and LVM setups. But it it’s a good procedure in applicable situations.

Show which process/program is listening to what port using netstat and lsof

lsof -Pnl +M -i4
lsof -Pnl +M -i6

or

netstat -tulpn
netstat -npl

There are obviously a number of ways to accomplish this, but these variations will cover a lot of ground. You can also combine this with grep to filter out things you don’t need to see, or to only include specific processes and/or ports.

See post from @geek1968 on Instagram

Apache goodies for WordPress security

The list of things to do to harden a WordPress site with Apache is long, but some things that could be done include:

FileETag None                                                                                                                       
                                                                                                                                    
<Files wp-config.php>                                                                                                               
    Require all denied                                                                                                              
</Files>                                                                                                                            
                                                                                                                                    
<Files xmlrpc.php>                                                                                                                  
    Require all denied                                                                                                              
</Files>                                                                                                                            
                                                                                                                                    
<LocationMatch "/wp-content/uploads/.*(?i)\.php$">                                                                                  
    Require all denied                                                                                                              
</LocationMatch>

SSH tunnel to use other mailserver than localhost

Because I have a lot of virtual machines, laptops, work environments, and so on, I never seem to find the time to setup SMTP authentication everywhere. I typically use Linux for everything except hardcore gaming, so it’s only natural that I have some sort of mail server installed like Postfix. The problem in using that mail server to send e-mail is that I also quite often have dynamic IP addresses on these machines, which doesn’t work well with “e-mail protection” (well..) like SPF.

So instead of making my life very complicated, I have a trusted server on the Internet through which I send e-mail.

If you were looking for something fancy in this article, you can move along now, there’s nothing to see 🙂

To make all my Linux work instances believe they’re talking to an SMTP server locally, I simply setup a tunnel from the given Linux instance to this trusted server on the Internet using the ever so versatile OpenSSH / SSH. I know there are a lot of ways to do this, but this is what works for me:

Local machine or “where I work”

I have a private/public key keypair on all of these machines. The public key is placed in the /root/.ssh/authorized_keys file on the trusted server that is running the mail server.

On this machine, as root, I setup a tunnel that looks like this:

ssh -N -L 25:localhost:25 root@mail.example.org -p 2222

This will create a tunnel from “localhost” port 25 (where I work) to “localhost” port 25 on mail.example.org. It will connect the end point of the tunnel to mail.example.org on port 2222. If the mail.example.org server is running an SSH server on its standard port (22), you can remove the “-p 2222” part.

Mail server

On this server, I only need to put the public key from the local machine “where I work” into /root/.ssh/authorized_keys to allow the tunnel to come up.

When I access port 25 on my local machine “where I work”, it will be sent through the tunnel and then attempt to access “localhost” port 25 on the mail server. The mail server software, Postfix in my case, will never know this connection did not actually originate from “inside” the machine, but rather through the tunnel.

Closing thoughts

You can (obviously) make this somewhat more automated with tools like AutoSSH, init scripts, and what not. The above only intends to show how uncomplicated it is to create useful SSH/SMTP tunnels 🙂