RE: creating complete OS image

["ed neville" <Edward@UKplc.net>]

R. W. Rodolico wrote:
> I generally just back up my data since it is pretty straight forward 
> to re-install the operating system. I point Apache DocumentRoot to 
> someplace in the /home directory (I do
> /home/http) and have a cron job back up the databases into 
> /home/dbbackup. Then, I have a cron job that rsync's /home, /root and 
> /etc to another machine each night.
> 
> As far as the operating system is concerned, there is a way to save 
> your package selection (ie, what packages are installed on a 
> particular machine) on one floppy.
> 
> Using this, I have recovered a dead machine in a few hours before.
> 
> Don't know if this is "best practise" but it works for my small 
> operation.
> 
> Rod
> 
> 
>> Hello People,
>> 
>> What's the best practise when doing disaster recovery ?
>> 
>> I have servers which I've configured putting quite some effort. I'm 
>> scared what would happen if the hard drive crashes. As a prevention 
>> measure I was thiking of making out a complete image of my present 
>> OS.
>> 
>> Now if the machine dies (say the hard drive) I could quickly replace 
>> the hard drive, re-deploy the image on it and restart the machine.
>> And the server should work absolutely fine.


Hello,

I'm new on this list.

The best thing I've found is before the box gets deployed take a dd
image of the disk using a file store box:

# dd if=/dev/hdd of=/images/big_router.img bs=8192

Then put the disk back into it's original box and then do nightly tar
balls:

# tar cjvf /root/backups/nightly/$( date ...) /etc/ /var/lib/mysql

Then scp that tarball to the filestore, and delete last weeks using ssh
root@filestore "rm /backups/date/big_router.tar.bz2"

Has anyone noticed weird behaviour with proftpd pam?

--
Regards,
Ed