so you have R2 and BR-BCVs presented to the same host ?

@Dynamox... Yes, we have the same host for both.

@Allen, thanks for your inputs, but our Team lead is hesitant to mask the R2's is there any other alternatives.

Thanks in advance

We do something similar using Clones instead of BCVs. For DR tests we want the DR hosts to come up seeing the Clones, but for a real DR we want them to see the R2s. So we normally leave the R2s masked to the hosts and the Clones unmasked. We have scripts that can quickly unmask the R2s and mask the clones for a DR test, and other scripts that reverse the process after the test.

This way the DR hosts are always ready for a quick DR activation.

Well, I may be misunderstanding, but in the situation you described you would normally not have the R2s masked in.

If for some reason you really don't want to play with the masking there is one other alternative that we discussed here at one point.

Put two Dual headed HBAs in your server. Have port 0 from Each HBA masked to the BCVs. Have Port 1 from each HBA masked to the R2s. That way, if you can normally keep the Port 0 connections active and work off your BCVs. In the event of a DR you would move the connections from Port 0 to Port 1 on each HBA then boot.

It would cost more for the HBAs, but you wouldn't need any more switch ports or FAs on the array.

Just a thought.

Another option .. map clones on a FA port and R2 devices on a different FA port.
Keep second FA port (showing R2 devices) offline. When you need to "see" R2 devices, put first FA port offiline and put online the second one.

Obviously you can't share the 2 FA ports with hosts having different requirements.

Message was edited by:
Stefano Del Corno

Thanks for all your answers, we will discuss on our Storage weekly meetings and will let you know what we decided.

I, we have exactly the setup you are describing: a DR site with R2s and Clones.
Clones used for DR testing and the R2 (synchronous) for real disasters.

Now the way we manage the assignation between R2 and Clones is simply via LUN Masking. Everything is scripted so it takes seconds to mask/unmask volumes to a specidied system if we need the R2 or clones.

So by default all systems are Offline. And based on what we need to do, the proper masking is applied before we boot the systems. So that way, systems are always booting with the right volumes. We usually unmask the clones when they need to be refreshed.

Depending on the OS, you may need a cleanup procedure or reboot when moving from R2 to Clones and vice versa.

Hope it helps

if you have access to SIU (Symmetrix Integration Utilities) you could use symntctl to control which volumes are mounted and which are not. When you unmount volumes it can put a lock on the volume so it does not get accidentally remounted. So you could keep both R2 and BRBCVs presented at the same time but only keep one set mounted. Check it out.

https://powerlink.emc.com/nsepn/webapps/btg548664833igtcuup4826/km/live1/en_US/Offering_Technical/Technical_Documentation/300-001-457_a05_elccnt_0.pdf

PowerPath is for path management, load balancing ..etc, SUI is more for logical volume management.

Thanks Dynamox, one question shall we use SIU if we are using Powerpath ? We have plans to use Power Path for all of our windows systems.

Normally I would agree with dynamox on using SIU to mount/unmount volumes, but in this case your initial post indicated you were concerned about the amount of time it took Windows to boot with all the devices visible at once. SIU won't help that. It will just change what file systems are mounted for the OS at any point in time.

good point Allen, SUI will not help with boot times.

@Allen Time taken is because it is mounting 300 devices instead of 150 devices, my main concern was We want to mount only BCV's and in real disaster R2 devices.


@dynamox
so SUI will not be useful to eliminate the 150 R2 devices during normal situation???

Nothing unsual in event viewer, my main concern was only the mounting of both the devices other than scripting shall we use SUI for that?

i am not sure if it's taking a long time during mounting or actually during the boot process where it has to scan every single device and probably not very happy to see 150 devices with identical disk signature. Anything interesting in Event viewer ?