32 of 34 people found the following review helpful:
2.0 out of 5 stars
Should focus on what we don't know, April 3, 2007
This review is from: Pro SQL Server 2005 Replication (Definitive Guide) (Hardcover)
I had a need... Bring up replication for a med-large Publishing DB (800GB).
I needed to configure, create test data, test several configurations, get a feel for performance and error handling, create the production scripts, prep the production environment, and turn it on.
The configuration we have setup involved "nosync" continuous Transactional Replication due to the Subscriber database starting out as an empty DB. The subscriber only receives new replicated transactions and does not have "snapshot" synchronization.
I had no real prior experience, nor guidance other than MSDN, MS Knowledgebase (KB), and various web sources.
Then, I came across this book, took a cursory glance, and saw what seemed to be a rich explication of replication. What caught my eye, in particular, were the chapters entitled: "Internals of Transactional Replication." There was such a chapter for each type of Replication (Snapshot, Transactional and Merge). I had hoped these would provide enough information to sort out problems I encountered and make my way through the task more efficiently.
Those were my expectations, and I was disappointed. Here's why:
1. The assumption is that Replication is going to use immediate Snapshot Synchronization. I got to believe there's a substantial number of DB's that would like to use it in a 'nosync' situation, particularly with a situation of Large DB replicating to Small DB. BOL, KB, nor this book really discusses this situation. There are specific and significant steps you must take in this situation. I was required to discover them through trial and error.
If you follow the directions given, you'll be forced to use Synchronization to initialize the Subscriber for Replication and this cannot happen in many situations.
Suggestion: provide a section for some types of "non-standard" database replication topologies and a summary of the differences. "Standard" topology, in this case, is where the Publisher and Subscriber are completely synchronized by an initial Snapshot Synchronization prior to replication transactions being distributed to the Subscriber.
2. This book is nearly 1000 pages. You'd expect a great deal of insight in that size of a book. I found little. What I did find was page after page of SQL Server Replication Wizard screenshots. I don't know about you, but I get pretty familiar with those wizards in the course of using them. They don't add very much value in a book like this.
Suggestion: Please don't provide out-of-the-box Wizard screenshots. If there is a situation that an admin will likely stumble over, then great! A snapshot of just that screen, pointing out what should be done is appreciated.
Suggestion: If you use diagrams (I love them!) provide more diagrams at a level like those in Chapter 3 that illustrate "how things work" by sequencing the replication types. Include some more non-standard ones perhaps in the Internals chapters.
3. The chapters entitled: "Internals of <Replication Type>" include ER-like figures. I say "ER-like" in that they are very inconsistent. In some cases there several Entity boxes (UML form) with the columns listed in a pattern in the figure. There are no relationship connectors, while it is clear there *are* relationships there. You're left to sort that out on your own. Effectively, what is added value is a listing of tables that are used for a specific form of replication. NOTE, *some* of the diagrams (Figure 10-29) have the relationships, but the Standard Transactional Publication figures (10-55 through 10-71) show almost no relationships. These were the ones I needed.
Suggestion: Create ER diagrams with the Entity names and their relationships (THESE ARE HELPFUL!). Don't try to list every column in the entity. If there are specific entities for which the column names are to be discussed, use separate figure(s).
4. Many of the non-wizard screenshots throughout chapter 10, the chapter I needed, were very blurry. I cannot even read some of them. I've looked at multiple copies of the book and they are the same. This is an issue with APress, not the author. This seems pretty unacceptable. Yuck.
Suggestion: proof *the entire book* before press...! :-)
5. Is it necessary (for even novices) to have 25% of the book showing Wizards and T-SQL for configuration? There are 6 chapters, 2 for each of the replication types entitled: "Configuring <Replication Type> Using the GUI" and "Configuring <Replication Type> Using T-SQL"
What's needed is:
- A conceptual understanding of what you're trying to achieve (terminology, roles, responsibilities, sequence of events) and I think you do this. This should provide enough information to identify the pattern of implementation you require.
- Next describe the procedures (tables, diagrams, enumerations) with the essence of how to implement the pattern. Look at the Table of Contents for "Configuring Transactional Replication using T-SQL" and you see the types of sequences that should be illustrated (not with Wizard GUI's). Then you have more room for non-standard configurations such as 'no-sync'. Specifically in this regard, nowhere is it discussed the correct procedure for generating the stored procedures and establishing them in the no-sync case.
6. Please focus on the "gotchas" what are the common mistakes you'll make and how do you rectify them. I believe the list is large and available, and it would be a great help to compile a succinct list/solutions.
7. More discussion is required regarding errors you may find in the Replication and what to do with them. There are a standard set of errors you may run into, and standard methods to deal with them, how to investigate MSrepl_errors, etc. You're not alone. There really is no great source for help in this regard.
Help other customers find the most helpful reviews
Was this review helpful to you? Yes
No
