![]() | |
![]() |
| | Thread Tools | Display Modes |
#31
| |||
| |||
|
|
In my organization we were thinking to use BIP IP for load balancing and in the database end we use Peer to Peer to replication so that BIG IP can redirect to any server based on less load and in the database end we have to make sure both the servers have up-todate data through replicaton and both are sync. Is it possible through P2P replication? real time replication and both the servers are active? and through BIP IP Hardware solution they can do load balancing. Thanks "Ekrem Önsoy" <ekrem (AT) compecta (DOT) com> wrote in message news:44240FED-D377-40CD-A604-3EEF76ACB6AF (AT) microsoft (DOT) com... Replication is not a "Hot Standby" solution, rather it' s a "Warm Standby" solution which means that you'll need to fail over to the secondary server manually. There are only two Hot Standby solution for SQL Server and they are Failover Clustering and Database Mirroring' s High Availibility mode. For example, for my last project I suggested to our customer to replicate their database to a secondary database and use this secondary database as a reporting server. In this project, I installed two clustered SQL Server intances on each server and this was a two-node cluster. On the first one, users perform OLTP and data on the first one is replicated to the secondary server using Transactional Replication and this secondary server is being used for reporting. We observed that a reporting query which runs on the primary runs for 258 seconds and when we run this very same query on the secondary node (the replicated one) it' s completed its execution in 128 seconds and this happens without any sql-tuning. This is all because of the new physical design. Because long running queries block and are blocked by other requests. This does not only slow down the reporting query but also blocks other insert\update\delete requests which is in nature of SQL Server in terms of ACID' s concurrency. -- Ekrem Önsoy "Rogers" <Rogers (AT) discussions (DOT) microsoft.com> wrote in message news:7A477C57-8B31-4AF6-AC04-90E289D95B56 (AT) microsoft (DOT) com... Thanks Ekrem, There are two purpose for these two way replication. 1. For Load Balancing 2. Failover. Thanks "Ekrem Önsoy" wrote: Failover Clustering is not a replication solution by itselt, it's used as a High Availibility solution. Database Mirroring must be the fastest one as a SQL Server solution however as you mentioned you can't use the Mirror database as a server just like the Principal. Do you want to use the second server as a reporting server or are you willing to manipulate data on that server too? If you mean to use the second server only for a reporting server then you can use Database Snapshots with Database Mirroring. So you'll be able to run SELECT queries against your Mirror database' s Snapshot. By the way, this feature can be used only for Enterprise Edition. If you want to manipulate data on the second server, then consider P2P Replication (http://msdn.microsoft.com/en-us/library/ms151196.aspx) Merge replication is more appropriate for "Server <-> Client" topologies... -- Ekrem Önsoy "Rogers" <Rogers (AT) discussions (DOT) microsoft.com> wrote in message news:AB0FB227-3798-416A-B760-E48C242D142A (AT) microsoft (DOT) com... Hello Guys, I would highly appreicate if any one can help me out on that... I was looking for Real Time two way replication but this should be 100% real time like we can't wait 3 seconds to replicate the data into another Server... I was using Database Mirrioring which is Active and Passive but I need Active / Active solution so we can use load balancing and also use for Disaster Recovery. Not sure about Peer to Peer Replication, will it faster than Merge Replication? Please adivce which one is the best soltuion is terms of very quick real time replication, I talked to couple of people and they were referring Clustering (hardwar solution) but can't we handle through Database Replication ? Please advice Thanks in advance. |
![]() |
| Thread Tools | |
| Display Modes | |
| |