Question? What is the best Disaster Recovery Plan?
Linux - ServerThis forum is for the discussion of Linux Software used in a server related context.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
Question? What is the best Disaster Recovery Plan?
This is a question out for all the RHEL and CentOS guru's and admins.
For weeks I've been trying out various Bare-Metal and Backup Rescue systems, and nothing seems to fit.
Here is my setup (basically):
CentOS 5.2
1 Tyan Tiger in Raid 5 - Mission Critical Production Server
1 Tyan Tiger unused clone of this system for parts
1 Tyan Tiger with 2 drives as raid 1 and one as tape storage for Bacula
4 HP ProLiant blades in raid 1 (mirror)- 1 mission critical
What worries me is, are these systems protected enough by the presence of RAID disks?
The Bacula server backs up all non-OS data daily from all critical machines.
Do I need a Bare-Metal Recovery plan on the raid 1 and raid 5 systems?
What would you (the other admins out there) do in this scenario?
Also, multiple disk failures are not unheard of (especially if the unit overheats for some reason) and can lead to total data loss despite RAID.
I backup to tape and keep a copy offsite for all critical production data.
Now, "bare-metal" backups are often silly because generally you have to install the operating system to restore the "bare-metal" backups anyway.
Back up your non-OS data and configuration settings, should be good.
RAID *helps* you to avoid needing to restore but it CAN fail so you always want to have a backup. I've seen two disks in RAID 5 die at same time which means the entire RAID 5 is dead. I've also seen bogus data go into the primary drive of a RAID 1 mirror. Since the data was at fault and not the hardware then the mirror drive also got the bogus data. In such a case you'd need to restore.
Most DR plans assume that not only your hardware but your entire data center may be kaput in a real "disaster". As such they usually not only have backing up the data but also sending it off site. Companies like Iron Mountain are around for storing your offsite data (and even picking it up if you're using tape).
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.