Migrate dead server to new.

[jeffdvo]

Hi All.
Bit of a forced situation, power supply on my (very) old Fujitsu Primergy Econel S2 (quad core 1GB RAM) server has died and it appears that rocking horse droppings are easier to find

Decided to replace server with a (used) Fujitsu Primergy TX140 S1p Tower Server i3-3220 3.3GHz 8GB, still awaiting delivery.

Looking for advice on how best to proceed in migrating old to new, with a dead old server! I've had a trawl through the boards and have found a few old posts on migrating to new hardware but haven't found any where the old server is dead and I'm not sure if they are relevant to CentOS 7.

Questions:-

1. Due to the difference in hardware specs, how likely is the "repair option" going to work?
2. If I do a clean install on "new" server, what files would I need to copy over from the old system disk ?
3. Is the "repair option" basically the same as running the OS from a USB stick?
4. Is Clonezilla (live) on a USB stick a good option for doing clones?

Other info that might be relevant! My knowledge of Linux is just enough to get me into trouble

The server is used at home as a SAMBA server. It mainly has just my wife and I accessing the net share. It has a physical "system" disk and 2 x 1TB physical data disks + a 4TB USB (backup) disk.

Any help or advice will be greatly appreciated.

Jeff

[TrevorH]

Theoretically it ought to be possible just to transplant the disk(s) from old to new system. Most likely it will fail to boot at that point but there should be a "Rescue" kernel listed as the last available one on the grub menu when you first boot up and that one should work on any hardware. Boot that and if successful then you need to regenerate the initramfs file for the most current kernel (or get the network working and `yum update kernel` if you do not currently have 3.10.0-957.21.3.el7.x86_64 installed).

First thing you need to know is the version of the latest kernel installed so run `rpm -q kernel` to list them. Pick the one with the highest version number and extract the version number from it - e.g if rpm -q kernel lists 'kernel-3.10.0-957.21.3.el7.x86_64' then its version number is '3.10.0-957.21.3.el7.x86_64'. If your currently installed kernel is not that one then it might be easiest just to yum update to that. If it is that or you cannot get the network to work then you can either `yum reinstall kernel-3.10.0-957.21.3.el7.x86_64` (also requires network) or you can run dracut -f --kver=3.10.0-957.21.3.el7.x86_64 to rebuild the current one.

Once you're booted up you may also need to fix up files in /etc/sysconfig/network-scripts/ifcfg-* to reflect new MAC addresses.

[jeffdvo]

Thanks Trevor, server coming on Monday!
Old server is as up-to-date as you can get, I think. I use Webmin and it lets me know about updates.
I can't really anything now until server comes but will let you know how things turn out.

Jeff

[jeffdvo]

Hi Trevor et all.

My new (to me) arrived late on Monday but it doesn't see my drives after completing bios. So I'm getting no where very fast
Server is Fujitsu Primergy TX140 S1p Tower Server i3-3220 3.3GHz 8GB Ram 3 x (hot swap) caddies. It seems is causing the problem, the Mobo has a raid controller stuck in a PCI slot and this controls the hot plug bays, never seen anything like it! Managed to find a couple of spare drives and played around with the raid config "WebBios" and ended up with a virtual drive!!! Anyway I was able to do an Centos (minimal) install on this and it seems to work OK.
From what I'm reading on Fujitsu forums etc, I might have to install each of my 3 disks as raid 0 and presumably have a virtual thingy with each one. A NAS box is looking very appealing! My old setup had a 160GB system drive and 2 x 1TB data drives + a USB 4GB backup drive.
Thinking the TX140 S1p was a bad purchase

If I ever get to booting from my old system disk and get into trouble, I'll give you a shout.

Jeff