TeeSeePeeEyePee
asked on
Off-Site Backup or Replication for a database on SCO Unix
I am looking for a solution to enable off-site backup, or ideally a redundant, replicated off-site server to provide fault tolerance for a database running under SCO Unix. The database is a proprietary one from a 3rd party vendor who does not offer any solution for this. If there is a way to replicate the entire server, that would be ideal, so that we could fail-over in the event of an outage. There is a point-to-point T1 available between both sites, and it should be more than enough bandwidth to handle replication in real-time or incremental nightly backups based on the amount of data change each day.
This project is in conjunction with similar off-site backup/replication implementation for Windows 2003 and Exchange 2003 servers, for which we have already found a solution, but if there is an application which can handle all three in a Windows environment, we would defintiely be interested in that as well.
This project is in conjunction with similar off-site backup/replication implementation for Windows 2003 and Exchange 2003 servers, for which we have already found a solution, but if there is an application which can handle all three in a Windows environment, we would defintiely be interested in that as well.
ASKER
Cyber,
Thanks for the ideas, I am taking a look at them. Basically the aim here is to not have to change the running server much (if at all). If there is an option, I would like to see something for UNIX like DoubleTake or NovaStor which we use for Windows servers. Along the same lines, we could build out clustering and AD replication on the Windows side, but these products make it a lot easier to configure, monitor and manage.
Thanks for the ideas, I am taking a look at them. Basically the aim here is to not have to change the running server much (if at all). If there is an option, I would like to see something for UNIX like DoubleTake or NovaStor which we use for Windows servers. Along the same lines, we could build out clustering and AD replication on the Windows side, but these products make it a lot easier to configure, monitor and manage.
I Understand that; But any layer you will add to get to your goal will eventually reduce performance. You may consider MS Storage Server which supports UNIX as well. But than again; you are not using any of the server natural environment where writing a script may contribute in both performance and reliability.
If you are intrested in the MS oriented solution I will post it to you.
Cyber
If you are intrested in the MS oriented solution I will post it to you.
Cyber
ASKER
We do not have the UNIX skills in-house to "home-roll" and support such a soultion, so if you have some ideas that are more MS-oriented, we could probably implement them far more quickly.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Those solutions we sell for lots of $$$$$.... why? Because youed have to analyse the way traffic floats and how to use this analysis to broadcast data into another and additional location and use that additional location to parallelicly calculate that data.
;)
But heres a tip:
UNIX has the ability to use shared computing architecture thus creating true cluster (if i'm not mistaken, you may use the 'split' or 'fork' commands to initiate such process). I am less familliar with SCO Unix but, sience it is a UNIX system, there might be a chance for you to do the same.
Links:
What is the 'Fork' handling?
http://web.ccr.jussieu.fr/ccr/Documentation/Calcul/usr-share/html/idebug/en_US/concepts/cbwfork.htm
Clustering architecture review in Linux (which may implcate the SCO UNIX as well):
http://lcic.org/reviews/0672323680.html
PS
Once you created clusterring; there is no problem in relocating the clusterred unit remotley...
Hope that helped in a way...
Cyber