Thanks everyone for your suggestions. This is all very helpful.<br><br><div><span class="gmail_quote">On 6/21/07, <b class="gmail_sendername">Steve Siegfried</b> <<a href="mailto:sos@zjod.net">sos@zjod.net</a>> wrote:
</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Michael Bruder asked:<br>><br>> Hi all,<br>><br>> Would any of you be willing to give me some advice on the best way to do a
<br>> bare metal recovery of Linux? The University that I work for is currently<br>> in the beginning stages of moving an Oracle 10g installation off of a<br>> dinosaur VMS server, and we cannot determine the best way to do a bare metal
<br>> recovery of the RedHat Enterprise 4 OS that will house the new Oracle<br>> installation. It was easy in VMS if we lost the operating system disk to<br>> just pop in a new drive, and restore the OS from tape. This doesn't seem so
<br>> easy with RedHat. We are currently researching options, but I would<br>> appreciate any input.<br>><br>> Thanks,<br>><br>> Mike<br><br>Create a "hot + spare" set of partitions and back up to 'em every evening
<br>using almost any of the "only copy the stuff that's changed and delete the<br>stuff that no longer exists" style backup programs _plus_ an extra step<br>at the end to reestablish the original /etc/fstab for the spare system
<br>(which will ensure the "spare" system can be booted as a replacement for<br>the "hot" one). The nice part about the "only what's changed" style<br>backups is that, even for very large systems, the backup usually runs
<br>in less than 20 minutes (and, if you're not overly picky about getting<br>every last user file updated, the backup can be run during "production"<br>time, too).<br><br>Important: When creating the "spare" partitions, make sure they're
<br>physically on different drives from the "hot" ones. That way, you can<br>lose a drive and still have an up-to-date bootable system.<br><br>When setting up grub, make sure you can boot either the hot or the<br>
spare system. If you lose the drive with grub on it, you'll need to<br>re-install grub, and after doing so, can re-establish /boot from the<br>backup copy.<br><br>I've been using this method for years, have had drives go bad and have
<br>never lost anything more than what changed since the previous backup.<br><br>As for the Oracle database itself, I'd suggest you use a mirrored RAID<br>array. It's not usually as fast as other types of RAID (e.g
.: striped),<br>but it is usually bulletproof with respect to having drives go sour.<br><br>Hope this helps and drop me a line if you need help finding the right<br>backup program'idly,<br><br>-S<br></blockquote></div>
<br>