HI,

I am a Linux administrator for a few months, where my network connects to another continent and I do not have any backup from the previous administration. Is it possible to make a server disk clone ??

Thanks in advance.

Welcome.

Some additional reading that may help:

• How to ask technical questions to get quality answers

Cloning may not be the best idea. There are other ways to make backups. What do you have in regards to distro, including version? Or is it a secret?

1 members found this post helpful.

https://wiki.archlinux.org/index.php/disk_cloning may help.

Thanks Turbocapitalist

Distro: Fedora 20 x86

- Have a DNS and DHCP server
- The Firewall
- Two Virtual Machines

What you suggest?

Upgrading. Seriously, Fedora 20 is ancient at this point..went end of life in 2015. Since you're the administrator, you should be asking yourself what you can do to ensure the SERVICES, not the machines, are available. A server is a server...they're interchangeable. One DNS server configured the same way, will provide DNS in the same manner as another.

Back up your CONFIGS, and test bringing them up on the latest version of whatever OS you're going to use. F20 needs to be put away, and if this is a real SERVER, then it needs to be running a server-class OS, like CentOS or Red Hat (if you're going to PAY for RHEL). The time to test your migration/DR plan is while your original server is available, not after the fact.

Your virtual machine disk images are just files...copy them somewhere.

1 members found this post helpful.

Linux distro's like Fedora tend to have software packages to assist users. Various forms of backup/clone/copy and syncing are available directly. It is agreed that best practices require updates and backups.

I would be tempted to look at rsync myself. Not sure remote clone is a great idea. Problem here may be live state data issue. Only one company that I know of has live state backup but I think I heard about an open source if you can find it. Various programs running on a server may not copy over properly by the way when running. Databases are terrible at this.

This is what a maintenance window is for. Shut down the service, copy/clone, locally if you have to, then restart the service.

You can't always use shutdown, nice but may not be handy.
Windows users enjoy live state backups so why doesn't linux?

I've never had a problem resurrecting a server from a failure with *nix, and can't say the same for Windows.

The services on *nix typically employ a config file of some sort (smb.conf for Samba, etc.)...and I can copy those easily while the server is running. A DHCP leases file won't matter. DNS/Bind? Websites (PHP/HTML?)...all can be just copied. Databases? Snapshot via mysqldump, RMAN for Oracle, etc. Bring a new server up with those same configs, and things will 'just work'. Same for things like webmin file(s)/'databases' for users and whatnot.

Typically, when I'm building out new server(s), I'll put something into a text-file as I go. If I modify a config/whatever, I'll put a reference into that file. Then, I'll use rsync/whatever to copy those file(s) to another location/server/disk/tape. That sever goes casters-up? No worries...quick reload from original install media, copy those files back, restart services. Back to square one.

It's a pain to keep that file up-to-date, but well worth it. If I make symlinks or other such things, I note those in that file too. Recovery time is VASTLY shortened.