Ferguson (4/30/2011)

---

Just be aware that IF UPDATE does NOT check if the value is changed, only if it is updated (i.e. in an update statement). So if you had...

STherrien-978820 (4/29/2011)

---

The main limitations I'm seeing that might be an issue:

1. I can't refactor tables too much, because splitting one table into three I can use a View to abstract...

This may be of use also.

select getdate(), [object_name],[counter_name],[instance_name],[cntr_value],[cntr_type]

from sys.dm_os_performance_counters

where [object_name] LIKE '%Replication%'

That column is added by replication when it does not have a unique identifier in a table.

Hopefully this will get you started. Run it in the context of the...

Stefan Krzywicki (4/28/2011)

---

If and only if they are generating new uniqueidentifier values for these inserts. If there are existing ids, those...

I like the following.

select floor(convert(decimal(15,8),100.988907))

, convert(decimal(15,8),100.988907)

, ceiling(convert(decimal(15,8),100.988907))

Alternatively, to reduce index fragmentation you could use NEWSEQUENTIALID()

I would use two temporary tables and insert into ... exec proc for both. This would give you the opportunity to tune the resultant query by apply indexes...

It is very easy to restore databases and to move the files in one step.

The restore statment includes the MOVE syntax which takes the logical file name and the new...

I'm a little puzzled as to why you would have a four node cluster and utilise only two of them? Wouldnt it make more sense to have two, two...

It will result in a connection failure and the query will stop running as part of the controlled shutdown of services.

All services are shutdown on one node and restarted on...

Assuming you are using windows authentication this is a pretty efficient way of doing it and takes away any requirement for developers to "remember" or potentially manipulate their code.

The storage needs to be active and presented to all nodes on a cluster. You dont need to make the disk available to windows as a file system drive...

Both the individual nodes and each sqlserver instance clustergroup require their own IP addresses and entries in DNS. I would look at your domains DNS server for the three...

There are a couple of KB articles related to these symptom's. I had similar symptoms on an AMD cpu cluster and you can see the conversation in the below...