Navision and Database mirroring

mconnolly

Does anybody know where i can find white papers or other documentation on Navision and Database mirroring.
I'm trying to find out how Automatic failover using a witness server works with the Navision Client.

Thanks in advance.

Mark

Nobody

NAV can not take advantage of the automatic failover feature of DB mirroring because NAV does not use the SQL Native Client ODBC driver which is required, it still uses the SQL 2000 ODBC driver. I tried to force NAV to use the new SQL Native Client but it does not work, or atleast I could not get it to work. :-)

Savatage

and here's a good post on mirroring - not Native but anyway...
http://www.mibuso.com/forum/viewtopic.php?t=15964

mconnolly

thanks for your help guys.

On a similar note, what are the recommended methods of fault tolerence for navision databases.

we are also starting to see issues with autogrom writing to the log file on a cluster server.

Thanks

bbrown

The fault tolerance method selected is a matter of what you are trying to protect against. The three common methods (log ship, mirror, cluster) all have distinct advantages/disadvantages that must be weight for each possible application. In some situation it you may need to combine methods to achieve the desired level of fault tolerance.

This, of course, doesn't even start to approach the third-party solutions out there.

mconnolly

its basically so we can offer a prefered service for our clients to protect their Navision.

It seems that log shipping is prefered to mirroring, but in the last day or two we've encountered issues with disk performance in a cluster (to be specific, log file autogrow)

I guess the answer i wanted was "In a navision environment, the best fault tolerence solution for Navision is scenario A = xxxxx, scenario B = xxxxx" etc etc but i guess it isnt as easy as that.

Savatage

mconnolly wrote:

in the last day or two we've encountered issues with disk performance in a cluster (to be specific, log file autogrow)

a tad off topic but here's some thoughts on Autogrow
http://www.mibuso.com/forum/viewtopic.php?t=25198

DenSter

mconnolly wrote:

we've encountered issues with disk performance in a cluster (to be specific, log file autogrow)

First of all, you can make a general recommendation about fault tolerance and log shipping, but if there is no budget for the additional hardware, the customer's not going to do it. I do think it would be a very good service to your customers if you had something prepared that explains some of their options, with pros and cons.

Now about the autogrow. You should set the size of your log file to be big enough to hold all transactions between regular backup points. Then just leave the file alone, don't shrink it just leave it as it is and you should be all set. By shrinking the log file, and letting it auto expand, you end up with a severely fragmented file.