Tag Info

Hot answers tagged

1

The simple answer is No... You do not need to disable access to the system etc. Here is why and how it works... When you setup replication, the publisher creates a Snapshot of the data at that very moment you create the publication. There is also a Log Reader that is running at the same time with Transactional Replication that is catching EVERY Transaction ...


1

The table modification that you are trying to do was most probably done in the graphical table designer in SSMS. The implementation of table changes through the graphical designer is poor. When you save the changes, SSMS runs off and copies the entire table into a new table with the new design incorporated, then drops the 'old' table and renames the new ...


1

Will, I can see why you'd have this concern but you shouldn't need to worry about it. The load on SQLA should not change in this scenario. The Log Reader Agent is running on a third machine correct? You have SQLA (Publisher) running, a (Distributor) running Log Reader Agent and the (Subscriber) SQLB receiving the updates. SQLA will still write the exact ...


1

It is best practice to move the distributor off of the publisher to another SQL Server instance to avoid these problems. Creating the snapshot will put load on the system as it needs to BCP the data out to files. Are you trying to replicate the entire 100 Gig database, or just part of it? With 100 Gig database there will be a hell of a lot of IO being ...


1

Check whether Browser services is running all SQL instance. Are you using a named instance? Also see if there are an resource jobs are running.


1

If you’ve validated the permissions as it sounds like you have, the cause could the performing of an operation on a non-existing object. To identify which object, run a profiler trace while reproducing the error. Look for “could not find object” error text so you can identify which object is missing. For example, a sp_MSrepl procedure could be missing.


1

I solved it, due to the sql 2000 compatibility (80) of the database, I've set an option on the replication to: Convert MAX data types to NText and IMAGE. so by doing a convert to varchar(max) it was solved.


1

Yes there is. The way I do it is I have a SQL Agent job post tracer tokens at a regular interval to the publishers and then monitor their progress from the distributor (where their history is stored). I created a view in the distribution database that helps with that. USE [distribution] GO IF EXISTS (SELECT * FROM sys.views WHERE object_id = ...


1

I would say to look into SQL server transactional replication with immediately updating subscribers.


1

Depends what you nuderstand under "same physical host". YOu could: Run into VM limits (Hyper-V: 4 virtual processors per VM at the moment, which may max out) Run into IO limits and have different discs for the second VM. Otherwise you are right. Can only be a small installation anyway.


1

If your reports could be run in such a way as to block your application, then this is a no-brainer for me.


1

alter database [publisher] modify file (name = 'logical_name', filename = 'new_path') alter database [publisher] set offline physically move the files from the old location to the new alter database [publisher] set online Keep in mind that the logreader agent will hold a connection open to the publisher. You can stop the agent w/o breaking anything. ...


1

You need to make sure there are no other active SQL sessions to the publisher database or setting it to offline will fail. -GFT


1

Can you consider an ALwaysOn group with reado only secondaries? http://msdn.microsoft.com/en-us/library/ff878253.aspx explains it. That is as good as it gets - take the reporting out of the transactional database. Btw., 10 million rows is not a lot. This is a very small daily upload volume - I suggest to check your programming for obvious snafus.


1

The log backup will not free up the log for reuse. Transactions in the transaction log in the publication database that have been marked for replication will not be unmarked until they have been written successfully to the distribution database. Only then can that portion of the log be marked for reuse. From the moment you stop the Log Reader Agent you ...



Only top voted, non community-wiki answers of a minimum length are eligible