Full Server Rsync Migrations

There are times when you need to do a one to one server migration.  This can seem like a daunting task, however with the magic of rsync, performing full server migrations can be a fairly painless task requiring little downtime.

Prerequisites

On both the old and new servers, you want to ensure the following requirements are met:

1. Confirm both the old and new servers are using the same hardware architecture. You cannot perform an rsync migration if one server is 32-bit, and the other is a 64-bit system. This can be verified by running the following command, and checking to see if it has “i386”, which means 32-bit, or if both have “x86_64”, which stands for 64-bit.

uname -a
Linux demo 2.6.18-308.el5xen #1 SMP Tue Feb 21 20:47:10 EST 2012 x86_64 x86_64 x86_64 GNU/Linux

So in our example, I have verified that both the old and new servers are 64-bit.

2. Confirm they are both running the same exact version of the operating system. Normally this means simply confirming both servers are at the latest patch level which you can do by running:

yum update
cat /etc/redhat-release
CentOS release 5.8 (Final)

For this article, I will be using two servers that are running CentOS 5.8 (Final).

3. Confirm rsync is installed on both servers

yum install rsync

4. Clean up server before migration. Depending on the amount of files on your server, the initial rsync migration can take quite a while. So you will want to browse through your old server and remove any extraneous temporary, cache, or other large files that you no longer need. You should also check your logs and ensure that their sizes are reasonable, and archive or delete the older logs you no longer need.

5. If you are not going to be swapping IP’s, and simply updating DNS to point to the new server, confirm that all your domains on the old server have a very low TTL set in the zonefile. A TTL of 300 is usually considered the lowest acceptable TTL to set.

Begin server migration

The procedure I’ll be writing out below is a two step process. It is meant to help minimize the amount of downtime that is involved when you swap the IP’s or update DNS, assuming you have a low TTL set. The steps are below:
1. Perform initial rsync
2. Perform final rsync and ip swap (or DNS update)

The initial rsync is just used to get the majority of the static files over to the new server. The final rsync is meant to update anything that is dynamic, such as logs, updated web content, databases, etc.

So before we begin, you will want to create an excludes file on the old server. This file will tell rsync NOT to copy over system specific information that is not needed for your new system.

vi /root/rsync-exclude.txt
/boot
/proc
/sys
/tmp
/dev
/var/lock
/etc/fstab
/etc/mdadm.conf
/etc/mtab
/etc/resolv.conf
/etc/conf.d/net
/etc/network/interfaces
/etc/networks
/etc/sysconfig/network*
/etc/sysconfig/hwconf
/etc/sysconfig/ip6tables-config
/etc/sysconfig/kernel
/etc/hostname
/etc/HOSTNAME
/etc/hosts
/etc/modprobe*
/etc/modules
/etc/udev
/net
/lib/modules
/etc/rc.conf
/lock
/etc/ssh/ssh_host_rsa_key.pub
/etc/ssh/ssh_host_rsa_key
/etc/ssh/ssh_host_dsa_key.pub
/etc/ssh/ssh_host_dsa_key
/etc/network.d
/etc/network/*
/etc/machine-id
/usr/share/nova-agent*
/usr/sbin/nova-agent*
/etc/rc.d/init.d/nova-agent*
/etc/init.d/nova-agent*
/etc/rackspace-monitoring-agent*
/etc/rackspace
/etc/driveclient/bootstrap.json

The example above should cover a couple of different distros. But always review someone else’s example before applying it to your own systems!!

Now that we have the proper excludes in place, lets run a dry run of rsync to see what would have happened before we actually do it. Please note that this is to run on the old server. Replace xxx.xxx.xxx.xxx with the IP of the new server:

rsync -azPx -e ssh --dry-run -azPx --delete-after --exclude-from="/root/rsync-exclude.txt" / [email protected]:/

If all looks well, lets perform the actual initial rsync:

rsync -azPx -e ssh -azPx --delete-after --exclude-from="/root/rsync-exclude.txt" / [email protected]:/

Depending on how much data you have, this could take a few minutes, or many hours. Once this is complete, you will want to schedule a maintenance window to perform the final rsync and IP swap (or DNS update). You want to perform this during a maintenance window as you will need to stop any database services or anything else that has very dynamic data to avoid corruption. So in the example, I just have a basic LAMP server, so I will need to shut down MySQL before performing the final rsync. Here are the steps I’ll be using:
1. Stop MySQL on old server
2. Perform final rsync
3. On new server, reboot server and test everything
4. Swap IP from old server to new, or update your DNS accordingly.

On the old server:

service mysql stop
rsync -azPx -e ssh -azPx --delete-after --exclude-from="/root/rsync-exclude.txt" / [email protected]:/

Now we are ready to start testing our new server.

Testing And Go-Live

Lets wave that dead chicken over the alter, its time to see if your new server survives a reboot, and if all the services come back online properly. There are no guides for this. Just reboot your server, then test out your sites, databases, and keep a wary eye on your system logs.

Once you have confirmed that everything looks good, it will then be safe to swap the IP’s, or update DNS accordingly. In the event that a problem surfaces shortly after the migration, you always have the option of rolling back to your older server, assuming you won’t be losing any critical transactions.

Rsync Migration Guidelines

There are numerous gotcha’s and things that you must be aware of before you can confidently perform a migration. Migrations go bad all the time. There is no guranteed way of knowing when it will fail. But there are ways to minimize the potential for some problem creeping up a month after the migration. Outlined below are steps that should be taken before proceeding with the migration.

1. Evaluate the server to be migrated. This involves:
– How large are the drives?
– Are there any directories that have hundreds of thousands of files?
– Are there any directories that contain thousands of other directories?
– Is the server extremely busy?

2. Check your backups to ensure everything is in place
You must first determine how you are going to do the migration. Are you build a new server, then use the backups to do a server restore to it? Or are you just going to perform a straight rsync migration. If your going to attempt to utilize your system’s backups:
– Check to see when the last known good backup was. You need confirmation that its good, and also attempt to get a rough eta on how long a server restore will take.
– Setup a new server with the same EXACT specs as the orginal server.

3. Ask questions
Below are some basic questions that should be asked before performing a migration. It may help shed insight on the server’s day to day tasks to ensure a smooth migration. Things to ask:
– Are their any known quirks experienced from time to time on the server?
– What are the key critical services that need special attention when migrating?
– How can you test the server to ensure the migration was successful? ie. websites that can check that utilizies both apache and mysql
– When is a good time the server can be shutdown the services on the production box so the final rsync can be performed?

4. Perform phase 1 of 2 of the rsync migration
The goal here is to create a base system on the new server. You want to be able to get the majority of the data copied over. This is to minimize the downtime the public will have during the final rsync phase. There is no need to schedule this, this is safe to do whenever.
– If you are utilizing your backups, get the new server jumped, throw on a temp ip, and do a full server restore.
– If you are just going to use rsync for everything, be aware that the server may seem sluggish as rsync may eat up the system resources.

To perform the rsync, log onto the old server (the one currently in production), and start a screen session:

screen -S migrations

Create a shared key between the old server and the new server

cd / && exec bash
for i in `ls |grep -v 'proc|etc'` do rsync -axvz --delete-after -e ssh $i [email protected]:/; done

To disconnect from the screen session, just hit:

ctrl-a then hit d

Depending on how much data and what the data structure is, at best your looking at 3-5G per hour. To speed up, use a cross connect on the gig ethernet ports.

5. Final rsync

At specified time stop all services on the production machine except sshd. (its better to just drop to single user mode with networking). Then using the same screen session, type:

for i in s | grep -v proc' do rsync -axvz --delete-after -e ssh $i roo[email protected]:/; done

Once complete, reboot and wave dead chicken over alter. You will want to swap the ips after you verified the new server at least boots.

6. Testing

This involves:
– Confirm websites work properly
– Confirm you can send and receive email
– Confirm mysql is functioning
– Go through error logs and correct any problems.

7. Troubleshooting

If machine doesn’t boot, you may have to fix grub (redhat). Also make sure /etc/fstab and /boot/grub/grub.conf have the labels setup right, or just specify the device: ex. /dev/hda1

FreeBSD dump/restore migration

Utilizing dump/restore to migrate data to a larger drive is one of the easiest and fool proof methods to perform a hard drive upgrade in FreeBSD.

This guide is assuming that you have a vanilla FreeBSD 4.11 install on the primary drive, and a secondary drive setup with the label as /backdrive. This logic should apply with FreeBSD 4, 5, 6, 7.

Here is how we do this:

First, lets just do this step by step, so lets dump the / partition. This is to be used per partition, so when you do df, and you see multiple partition, do this for each partition. But as this server only has 1 partition, simply do:

dump 0f - / > root.dmp

Transfer over the file to the new server:

scp root.dmp [email protected]:/tmp

Next, on the new server, cd into the directory we’ll be restoring, and begin the restore:

cd /backdrive
restore -rf /tmp/root.dmp

Fix the disk label and make /backdrive the / partition to boot:

# Notice we're modifying this for the second disk
disklabel -e -f ad1s1

# Change:
disk: ad1s1 to ad0s1

# And change:
e: 154199170 2097152 4.2BSD 2048 16384 89 # (Cyl. 130*- 9728*)

# To:
a: 154199170 2097152 4.2BSD 2048 16384 89 # (Cyl. 130*- 9728*)

Then fix fstab to point to /dev/ad0s1a for /.

Extended notes on disklabel:
Ok, there is a small science with disklabel. But just watch to 2 things:

1. The e: is the root partition on the second drive. So in disklabel:

change e: to a: (as a: is the default root partition on freebsd). 

2. Now you have to match up whats showing in disk label to what is showing in fstab:

Before:
8 partitions: <-- (figure out this nightmare)
# size offset fstype [fsize bsize bps/cpg]
b: 4194304 0 swap # (Cyl. 0 - 261*)
c: 71119692 0 unused 0 0 # (Cyl. 0 - 4426*)
e: 262144 4194304 4.2BSD 2048 16384 94 # (Cyl. 261*- 277*)
f: 524288 4456448 4.2BSD 2048 16384 94 # (Cyl. 277*- 310*)
g: 524288 4980736 4.2BSD 2048 16384 94 # (Cyl. 310*- 342*)
h: 65614668 5505024 4.2BSD 2048 16384 89 # (Cyl. 342*- 4426*)

After:
8 partitions:
# size offset fstype [fsize bsize bps/cpg]
b: 4194304 0 swap # (Cyl. 0 - 261*)
c: 71119692 0 unused 0 0 # (Cyl. 0 - 4426*) 
a: 262144 4194304 4.2BSD 2048 16384 94 # (Cyl. 261*- 277*)
f: 524288 4456448 4.2BSD 2048 16384 94 # (Cyl. 277*- 310*)
g: 524288 4980736 4.2BSD 2048 16384 94 # (Cyl. 310*- 342*)
h: 65614668 5505024 4.2BSD 2048 16384 89 # (Cyl. 342*- 4426*)

And with fstab, I just have to verify each label by testing it and figuring out which is which.