Transcript Ch12-short
SQL Server 2005 Implementation and Maintenance Chapter 12: Achieving High Availability Through Replication Introduction • Use replication to copy data to different locations in your enterprise • There are several reasons – Move data closer to the users – To reduce locking conflictsTo allow site autonomy – To remove the impact of readintensive operations © Wiley Inc. 2006. All Rights Reserved. Publisher/Subscriber Metaphor • There are several key terms used in replication – Publisher • the source database where replication begins. It makes data available for replication – Subscriber • the destination database where replication ends – Distributor • the intermediary between the publisher and subscriber. It receives published transactions or snapshots and then stores and forwards these publications to the subscribers – Publication • is the storage container for different articles. A subscriber can • subscribe to an individual article or an entire publication. – Article • the data, transactions, or stored procedures that are stored within a publication. This is the actual information that is going to be replicated. – Two-phase commit • is a form of replication • in which modifications made to the publishing database are made at the subscription • database at the same time © Wiley Inc. 2006. All Rights Reserved. Articles • An article is data in a table • It can be the whole table or just a subset – Horizontal partitioning • Rows are filtered out – Vertical partitioning • Columns are filtered out © Wiley Inc. 2006. All Rights Reserved. Publications • A publication is a logical collection of articles • Subscribers subscribe to a publication – They do not need to read all of the articles though © Wiley Inc. 2006. All Rights Reserved. Replication Factors • Factors that affect replication are – Autonomy • Level of subscriber independence – Will the replicated data be considered readonly? How long will the data at a subscriber be valid? How often do you need to connect to the distributor to download more data? – Latency • Frequency of data updates – Transactional Consistency • Do all the transactions that are stored need to be applied at the same time and in order? What happens if there is a delay in the processing? © Wiley Inc. 2006. All Rights Reserved. Using Transactional Replication • Transactions from the publisher are stored on the distributor • Subscribers receive and apply transactions – They should treat data as read-only • Some latency and autonomy can be introduced • Subscribers do not need to be in contact with publishers at all times © Wiley Inc. 2006. All Rights Reserved. Using Snapshot Replication • Moves an entire copy of the published items © Wiley Inc. 2006. All Rights Reserved. Snapshot With Updating Subscribers • Initial replication works just like snapshot replication • Subscribers use 2PC to update the publisher • Publishers must approve subscriber updates © Wiley Inc. 2006. All Rights Reserved. Using Merge Replication • This allows subscribers to make changes to their local data • All changes are merged with all other subscribers – When all subscribers have the same data, they are in convergence © Wiley Inc. 2006. All Rights Reserved. Using Queued Updating • Used with transactional and snapshot replication • Allows subscribers to update publishers • Updates are queued, not immediate • There are other Restrictions © Wiley Inc. 2006. All Rights Reserved. Replication Internals • There are two types of subscriptions • Push – Configured and maintained at the publisher • Pull – Configured and maintained at the subscriber © Wiley Inc. 2006. All Rights Reserved. Publication Issues • Timestamp datatype – This is replicated as binary data in transactional and snapshot – Data is not replicated in merge • Identity values – You can assign a range of values to replicate • User-defined datatypes – These must be created on each subscriber • Not for replication – Prevents objects from being replicated © Wiley Inc. 2006. All Rights Reserved. Publication Restrictions • Replicated tables must have a primary key. – Except in snapshot replication. • Publications cannot span multiple databases. • Varchar(max), nvarchar(max), and varbinary(max) data is not replicated in transactional or merge replication. – Because of their size, these objects must be refreshed by running a snapshot. – Only the 16-byte pointer to their storage location within the publishing database is replicated. • You cannot replicate from the master, model, MSDB, or tempdb databases. © Wiley Inc. 2006. All Rights Reserved. Distributor Issues • • • • • • • • Ensure you have enough hard disk space for the Distribution working folder and the distribution database. Do not let the distribution database’s transaction log fill up. The distribution database will store all transactions from the publisher to the subscriber. It will also track when those transactions were applied. Snapshots and merge data are stored in the Distribution working folder. Be aware of the size and number of articles being published. Text, ntext, and image datatypes are replicated only when you use a snapshot. A higher degree of latency can significantly increase your storage space requirements. Know how many transactions per synchronization cycle there are. © Wiley Inc. 2006. All Rights Reserved. Replication Backup • There are four possible strategies – Publisher only – Publisher and distributor – Publisher and subscriber(s) – Publisher, distributor and subscriber(s) © Wiley Inc. 2006. All Rights Reserved. Replication Scripts • Use scripts to track different versions of your replication implementation. • Use scripts to create additional subscribers and publishers with the same basic options. • You can quickly customize your environment by modifying the script and then rerunning it. • Use scripts as part of your database recovery process. © Wiley Inc. 2006. All Rights Reserved.