Koozali.org: home of the SME Server
Legacy Forums => Experienced User Forum => Topic started by: Nate on November 15, 2002, 06:33:09 PM
-
My e-smith is modified heavily, so installing new then running a restore was going to take a long time. I tried to use partition magic to copy my old hard drive to my new 80 gig drive. It was a no go--error messages. This morning I used Norton Ghost 2003 and it worked great!! The new hard drive is running perfect, no problems at all.
Open ghost
Ghost Advanced
Click Run Ghost Interactively
Computer reboots to Ghost program
Clone the drive
That's it.
By the way, the Western Digital 80 gig hard drive with 8 meg of cache is very nice.
As a backup, I cloned 2 of the 80 gig drives. If the one is use fails, I can install the back-up drive, do a restore, and I'm back in business--without having to reinstall SME and customize it heavily and then do the restore.
Nate
-
of course you have an alternative backup mechanism operating that makes sure your users' data is backed up every night ;-)
-
I use ghost to do the same thing. Once a week I ghost the whole server, it only takes me like 10 minutes.
On regular days I used the how-to that showed me how to use samba to map a remote share and backup to that remote share every night. Then I burn the backups to a CD-RW.
Then if my HD crashes I can just reghost a new one and then restore my backup from last night. I tested this and it works like a champ. I have even tested to see if I can move the reghosted HD to another CPU that is close but not the same as my current server and that works great too. So I have an extra CPU in case the one I am using dies.
-
Sounds similar to my daily backup. My goes to a windows 2000 share on my workstation's secondary drive. Every couple of weeks I clean out the back-up drive and burn a CD. You could put the ghost hard drive in about any computer and re-run the configuration program to set up the network cards and stuff.
-
Dear All
You all seem to be quite concerned about your data etc and making sure your system is up and running, but no-one mentioned RAID1 setups. If you are concerned about the reliability of your data wouldn't it make sense to have RAID1 software or hardware setup on your servers. I know this is not the same as an offsite backup (which is still necessary).
I'm interested to know if you are or are not using RAID1 for data reliability as well as doing backups of your system. In theory if one disk fails the other one keeps going and your data is still intact and live !! Just replace the faulty drive and resync the data with no or very little down time
Any comments ?
Regards
Ray Mitchell
-
Well the reason why I don't use raid1 is when I first set up the server I am talking about I was using e-smith 4 and I didn't set up raid. So I have to protect the configuration and data I have on there now.
Trust me if I could change it to raid1 I would. ;-)
-
Upgrading to Raid 1 can be done, but it's a bit tricky. I just did it a couple of weeks ago, changing from 11GB to 80GB in the process. Don't have a detailed howto, but here's the very short version from memory:
1. Remove original drive from system, install new drives as hda and hdc (CDROM is hdd).
2. Perform a clean install onto the new drives, using software RAID 1.
3. Jot down, print out, or otherwise make a record of the contents of the /etc/raidtab file.
4. Switch the cables around--move hda to hdc, hdc to hdd, and reinstall the original drive as hda.
5. Boot off the original drive, probably into single-user mode (I didn't do this, but it's probably a good idea).
6. Recreate /etc/raidtab, changing hda to hdc and hdc to hdd as appropriate.
7. mkdir /mnt/raid
8. raidstart /dev/md1
9. mount /dev/md1 /mnt/raid
10. cd /mnt/raid/etc; cp lilo.conf raidtab modules.conf /root
11. cd /
12. cp -ap bin home opt quota* root service usr etc lib sbin tmp var /mnt/raid/
13. mkdir /mnt/raid/mnt
14. cd /mnt/raid/mnt
15. mkdir floppy cdrom
16. cd /root; cp lilo.conf raidtab modules.conf /mnt/raid/etc.
17. Shut down
18. Remove the original drive, change hdc back to hda, and hdd back to hdc.
19. Reboot. Cross fingers.
Now, if you haven't installed any custom software, none of this is necessary. Just do a system backup, do a fresh install using software RAID, and then do a restore. Much simpler, but if you've installed any extras, you'll have to redo them.
-
I'm not concerned about being down a few hours, or however long it takes to install my back-up drive. My production SME 5.1.2 server is purely email. Our LAN file sharing server is a raid 5 configuration(SCSI drives) that contains many huge databases and lots of important user files that are need for our office to function. Some day, when I get tired of the NT4 file server I'll probable install an SME raid configuration on a separate file sharing server--maybe with some SCSI drives.
-
Dan,
Do you know if your method would work on scsi drives?