Friday, March 30, 2012

Replication Test\Check 6.5 & 2000 are in sync

Hi,
I am having MSSQL 6.5 with batch driven transactions
I am upgrading to 2000 and plan to do some black box
testing.
To test the batch jobs,I am planning to run the job on 6.5
and 2000 and test the rowcounts and records added/updated
based on datetime.(I don't have delete..so checking the
data on the basis of timestamp works)
However my rowcount and timestamp queries take a long time
to run on 6.5.(More than 50 mins as data volume and no of
tables are high)
I can't wait that long as the production server 6.5 has
jobs that run after every 30 mins.Is there a way just to
check whether 6.5 and 2000 data are in sync using some
built in utility or better algortithm(I don't want to
compare using SQL,but comparison at the lower level such
as file etc would do)
Basically do something like what replication or mirroring
does but not actually move the data but just check
Thanks
RakeshTake a look at the sp sp_replcounters... It shows how many records in the
log are awaiting replication...
"ar" <ambewadkarrakesh@.johndeere.com> wrote in message
news:05fb01c344f6$cb8265b0$a501280a@.phx.gbl...
> Hi,
> I am having MSSQL 6.5 with batch driven transactions
> I am upgrading to 2000 and plan to do some black box
> testing.
> To test the batch jobs,I am planning to run the job on 6.5
> and 2000 and test the rowcounts and records added/updated
> based on datetime.(I don't have delete..so checking the
> data on the basis of timestamp works)
> However my rowcount and timestamp queries take a long time
> to run on 6.5.(More than 50 mins as data volume and no of
> tables are high)
> I can't wait that long as the production server 6.5 has
> jobs that run after every 30 mins.Is there a way just to
> check whether 6.5 and 2000 data are in sync using some
> built in utility or better algortithm(I don't want to
> compare using SQL,but comparison at the lower level such
> as file etc would do)
> Basically do something like what replication or mirroring
> does but not actually move the data but just check
> Thanks
> Rakesh

No comments:

Post a Comment