SEARCH
0-9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Prev | Current Page 344 | Next

James Luetkehoelter

"Pro SQL Server Disaster Recovery"

We now had two instances of that
Java application sending duplicate records. IT staff on the business partner??™s end worked
with the data sent as best it could, but our data, which incidentally got replicated all over
the place after it was transmitted and consequently received back, was an unknown
amount of garbage. And that resulted in a manual cleanup that impacted us for weeks.
Needless to say, the configuration has been corrected since then, but this is an example
of how clustering is not the high-availability answer to everything.
For applications that are cluster-unaware, you must create manual compensation
steps to include the cluster-unaware applications in the entire clustering solution. In
the previous example, our solution was to have a simple batch job that fired whenever
a new instance of the Java application started. The role of this batch job was to see if an
instance of the Java application was running on the other server participating in the
cluster. The batch job was included in SQL Agent and set to fire whenever SQL Agent
started. If SQL Server failed over (and not the entire node), SQL Agent needed to fail
over as well (it relied on the MSDB database).


Pages:
332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356
hotel jelenia góra Russian bride Free English grammar and study guid powiekszenia wielkoformatowe counter strike 1.6