October 24, 2013 at 2:32 pm
Hi,
We have mirroring set up as part of DR solution. Since the database is highly used OLTP database, we dont really get a maintenance window to rebuild indexes. The fragmentation is at its high peak.
We thought of replacing mirroring with replication (peer - peer), so that, we can do maintenance in one publisher while the other publisher takes active transactions.
Now, I am aware replication is the right strategy for DR. However, if i go for this option, what will i miss? Mirroring mirrors what ever happens in the source. Is that the same for replication as well?
Any comments on this is highly appreciated.. Thanks
October 25, 2013 at 3:05 am
Mirroring is HA, replication is data distribution.
In P2P replication you have no primary/secondary or principal/mirror relationship between nodes.
I highly discourage that approach.
I would focus on the main problem: rebuilding indexes.
Why do you need to rebuild indexes in the first place? Reorganize will do most of the time.
Ola Hallengren has a very nice maintenance script that you could use to skip/reorganize/rebuild indexes based on their fragmentation level. You can find it here[/url].
The script detects which edition you're running and tries to rebuild/reorganize indexes online, which could save you lots of blocking.
Also, if you get highj fragmentation levels, your indexes are probably not optimal in terms of fillfactor or pad_index. Look into that first.
-- Gianluca Sartori
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply