Re: Image/backup

• Previous message: Krautkramer, John: "NFS Mount 'Permission denied'"
• Maybe in reply to: Andrew.Bridgeman_at_corusgroup.com: "Image/backup"
• Next in thread: Robert Williams: "RE: Image/backup"
• Reply: Robert Williams: "RE: Image/backup"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Date: Tue, 2 Aug 2005 10:51:28 -0700 (PDT)
To: redhat-list@redhat.com

This question is so basic, yet so complicated to
answer. The first question is, is "How critical is
this system?" Think about it. What if your Internet
connection goes down? What if you have a redundant
Internet connection and the city looses power?
Happens in Southern CA. could happen in your
town/city. So the first thought is should you have a
redundant system somewhere in case your location went
down? What about a natural disaster?
To answer this, if you need that kind of reliability,
I would suggest planning something out with a 24/7
Webhosting company that guarentees uptime and
redundant power. A small colo cabinet should be
enough for your needs. Although, even then, I would
have a redundant backup system at your business
location. I would also recommend that the colocation
facility be in another state. For even more
complicated and complete solutions, Akamai offers
more
http://www.akamai.com/en/html/services/enhanced_dns.html
I like the cluster idea, but I do not believe it
necessary. A Cluster is more for a high performance
environment. It will allow you more performance, but
for the time and the upkeep, it's just frustrating to
administrate.
I would advise using something like a high availablity
server: http://linux-ha.org/ I have used heartbeat in
a number of environments. In a Heartbeat environment
you have two servers. One is the primary and Two is
the backup. If for some reason One fails, then Two
assumes the ip and services of One. in essence it
becomes One. It does this by mirroring the content on
One, then in the event of a failure, which it checks
for, it takes over.
Your solution is weak. If the raid controller dies on
this server. Your down. If the power supply dies
(and redundant power supplies do fail) then your down.
 If any number of hardware issues occur, you are
down.
In regards to Disaster Recovery. Providing that you
have something like heartbeat running, in the advent
of a disaster, you will want something that can get
your primary server back online as soon as possible.
I would recommend buying 3 servers that have exactly
the same hardware. Should One or Two go down, you
could always bring up the third or use parts from the
third. I would recommend having a default image of
the OS stored somewhere with all the basic software
configured. Possibly on a fileserver or tape. Tape
will be slower to restore from. But, the image will
need to be maintained, and of course you will need a
backup system for the dynamic data. Whether it be
database or web site. In either case you should be
able to have weekly, and daily backups of that data
for at least 2 weeks if not more. There are a number
of commercial and open source products available to
meet your needs there.
My last bit of advice is, Test. Test. Test. Test
your backup disaster solution. Test it monthly. If
you can't find the time to test monthly, hire someone
to assist. It is absolutelty critical that you are
prepared. Just like the fire alarm tests, we all had
in grade school, you need to be able to get your
system back in the advent of a real disaster.

That's all I have. There are a number of good books
on the subject or disaster recovery, I suggest that
you think about reading one or two of them.

Jay

We currently have a Redhat machine that is a 24/7
machine and We currently have a Redhat machine that is
a 24/7 machine and is critical to our company. We want
to be able to get the machine back to its current
state if it where to blow up or something else extreme
should happen. Basically we want to be able put two
new disks in the machine in question in the event of a
problem and be up a running within minutes. The
current two disks are Hardware mirrored so we would
need the same info/data/config on two new
disks and then they can sit in our safe incase of a
emergency. One other thing to add is that ideally we
do not want to have any downtime on the server as i
said before it is used 24/7.
       
Could anyone please advise in detail on how we might
achieve this?
                                                   

__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

-- 
redhat-list mailing list
unsubscribe mailto:redhat-list-request@redhat.com?subject=unsubscribe
https://www.redhat.com/mailman/listinfo/redhat-list

• Previous message: Krautkramer, John: "NFS Mount 'Permission denied'"
• Maybe in reply to: Andrew.Bridgeman_at_corusgroup.com: "Image/backup"
• Next in thread: Robert Williams: "RE: Image/backup"
• Reply: Robert Williams: "RE: Image/backup"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]