April 28, 2009 at 4:48 am
Can someone give me a quick overview of obvious changes I should note?
It seems everything is timing out and I have to keep changing timeout settings. The monitoring is not as straight forward as 2000 (replication monitor is pants so far). It really is rubbish. Under sp_who2 I see 3 occurences of 1 log reader. Is this normal behaviour? one flits between Suspended and sleeping.
Our server ran out of space so the log reader stopped. Seemed to rollback first. So what state is it in? It gives no clues.
I feel as though its not so easy and manageable as 2000 and obvious as to the replications tatus of the data. Haven't a clue if it is all in sync.
If I kick off a snapshot sync will it duplicate all the data? Seems like it wants to.
Any help?
April 28, 2009 at 8:49 am
Shark Energy (4/28/2009)
Can someone give me a quick overview of obvious changes I should note?It seems everything is timing out and I have to keep changing timeout settings. The monitoring is not as straight forward as 2000 (replication monitor is pants so far). It really is rubbish. Under sp_who2 I see 3 occurences of 1 log reader. Is this normal behaviour? one flits between Suspended and sleeping.
Our server ran out of space so the log reader stopped. Seemed to rollback first. So what state is it in? It gives no clues.
I feel as though its not so easy and manageable as 2000 and obvious as to the replications tatus of the data. Haven't a clue if it is all in sync.
If I kick off a snapshot sync will it duplicate all the data? Seems like it wants to.
Any help?
I have to agree with you in many of your points.
Replication monitoring/management took definitely an (or many) step back from UI perspective.
Agents in SQL 2005 though are much more "resilient" to failures in terms of "re-trying" but the UI does NOT helps
setting up of replication is basically very similar but if you can stick to TSQL you will be much better of.
New things introduced that I have found valuable:
- DB Snapshot for concurrent processing ( this is available only on EE though)
- Initialization from database backups (time saver for large databases)
- DDL replication (this is a biggie for those shops where schema is changing frequently. Is there anyone who isn't? )
- Higher limits on columns per table, wider datatype range, a little higher performance
There are others that at least on 2005 are not so clearly useful at least that's my opinion.
* peer-to-peer (they actually got it right in SQL2008 but in 2005 it requires to quiesce the entire topology to update schema)
* granular security this also seems to me like an unfinished business you still need higher level permissions regardless of what the documentation says for certain opperations. Use always sa-level account if you want to avoid problems at setup time
* multiple subscriber streams. This was (not sure if it is still, I stopped trying) riddled with bugs!
etc...
Anyway, trying not to use the UI unless you have to, has been a great decision on my firm.
We have our own scripts to manage, monitor and set up the whole thing.
The UI is NOT your friend as in SQL 2000. Sadly but true!
* Noel
Viewing 2 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply