<p dir="ltr">Mail servers are hard to backup. I've setup nightly with hourly sync. That worked OK until it got too busy. What I eventually used was dual systems mirrored in DNS with split drive arrays in mirrors on each. A daily backup then was sufficient. They were running several hundred messages an hour. I imagine at hundreds per minute or higher this becomes no longer a backup issue but a purely continuity issue.</p>
<div class="gmail_quote">On Feb 12, 2014 4:00 PM, "John Heim" <<a href="mailto:john@johnheim.com">john@johnheim.com</a>> wrote:<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
On 02/12/14 10:38, Brian Mathis wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
<br>
<br>
'rsync' is still a good option to copy files back. It does not do a full check of files unless they already exist on the destination side, and even then it checks the date and filesize before doing a full checksum. If you use the options, you can preserve timestamps, view progress, and easily stop/restart the copy if needed. I'd use these options: -av --progress --partial. When copying between local disks, it's just a version of 'cp' that gives you a lot more options. Also, I'd recommend running this inside of a 'screen' or 'tmux' session so you can detach and leave it running, and/or protect against the session getting dropped from the network.<br>
<br>
As far as step 5 goes, do you have enough physical space to add the new drives next to the old ones? Then you could make a partition for the user data and stage it directly on the new drives.<br>
<br>
Incidentally, the point of a RAID is to be able to replace failing drives without having to go through this process. Is there any reason why you can't just replace the failing drive and rebuild the existing RAID to it?<br>
<br>
<br>
</blockquote>
<br>
We have one failed disk and one with failure predicted. This is a Dell machine and they recommend rebuilding the whole thing when that happens. There is something called "puncturing" which is actually a good thing. It allows the RAID5 array to keep running even if 1.5 disks have failed.<br>
<br>
We just got unlucky and had two disks fail at the same time. Well, only one has actually failed. The other just is in failure mode.<br>
P.S. RAID is not a backup and you should always have some kind of backup job running to external media regardless of the health of the RAID.<br>
<br>
Well, it's a mail server and nightly backups aren't good enough for this purpose. We don't want to restore from last night's backup and lose 8 - 10 hours of mail.<br>
<br>
______________________________<u></u>_________________<br>
Ale mailing list<br>
<a href="mailto:Ale@ale.org" target="_blank">Ale@ale.org</a><br>
<a href="http://mail.ale.org/mailman/listinfo/ale" target="_blank">http://mail.ale.org/mailman/<u></u>listinfo/ale</a><br>
See JOBS, ANNOUNCE and SCHOOLS lists at<br>
<a href="http://mail.ale.org/mailman/listinfo" target="_blank">http://mail.ale.org/mailman/<u></u>listinfo</a><br>
</blockquote></div>