Unsolved
This post is more than 5 years old
236 Posts
0
2084
RecoverPoint Lag
Hi All,
I was wondering if lag can only be controlled by configuring RPO?
Appreciate your input.
Unsolved
This post is more than 5 years old
236 Posts
0
2084
Hi All,
I was wondering if lag can only be controlled by configuring RPO?
Appreciate your input.
Top
forshr
1.1K Posts
0
June 2nd, 2015 12:00
It can. From version 4.0 the Lag Optimizer feature was introduced. You can find reference to this in the Admin Guide.
Regards,
Rich
victory_is_mine
236 Posts
0
June 3rd, 2015 06:00
As far as I understand the lag is measured in GB, isn't it? It tells us how much data has to be transferred to the copy volume to make it identical to production. How can this be translated into seconds, minutes, hours? I am trying to figure out the correlation between the two. In other words, if the lag is for example 2GB, does that mean that the copy volume is 2GB behind the production volume? How many seconds, minutes, hours, etc. is it behind, how do I figure this out?
forshr
1.1K Posts
0
June 3rd, 2015 07:00
The default is 25 seconds and I recommend you use time as the estimator for data lag. Using data as the estimator is not a normal requirement.
Regards,
Rich
victory_is_mine
236 Posts
0
June 3rd, 2015 13:00
OK, what does it actually mean if the RPO is 25 sec: is it that my copy and production volumes are going to be different by that amount of data that was written to the source in 25 seconds before disaster, and it was not transferred to the copy?
What is going to happen if the RPO is set to 20-30 minutes?
forshr
1.1K Posts
0
June 3rd, 2015 14:00
Correct, and more specifically it will try to meet this RPO as a minimum at all times but note that the Lag Optimizer has two aspects to it. This is one and the other is the Group Priority option. Take a look at the RP Admin Guide which explains the feature.
victory_is_mine
236 Posts
0
June 3rd, 2015 15:00
Great, thank you.
One more question: RecoverPoint journals do require dedicated RAID groups, preferably consisting of SAS drives, right? It is not the best practice to configure them on the pool LUNs, is it? I was looking for the EMC document to prove it, but cannot find anything.
forshr
1.1K Posts
0
June 3rd, 2015 16:00
Pools are absolutely fine. They have the advantage of offering easier management when it comes to provisioning LUNs for the CGs plus using the largest stripe size cross a pool dovetails well with the large sequential I/O nature of the distribution process from the RPAs. Just ensure they are sized correctly and homogenous but are not part of a FAST configuration or have FAST cache enabled. Also, thin LUNs for journal pools are pointless.