Migrate FC R6 CPG to NL

[ConnieM]

hi team,

I'm fairly new here,

1. Can you advice, I have CPG for email logs directories configured as follows, FC R6 with 28 VV, i want to create same CPG but now on NL and copy that CPG contents to new CPG, reason being I'm running out of space on FC disk? what is teh best practise? 

2. Basically VV are exported to host set1 that comprises of 14 VMs, another question is that why do I now have doubled VV, totalled to 28 VV?

[Cali]

Create the new NL CPG

Go to one existing Volume > Tune > to the new NL CPG

This migrate the Data inside the existing VV to the new CPG (Online).

I would do this Volume by Volume.

There is no need to create new NL Volumes.

[Sheldon Smith · Accepted solution]

Hi ConnieM,

There are two things to consider when adding volumes and moving volumes around on a 3PAR system. One obviously is capacity. You need the physical place to store the data. The other is performance (IOPS and Bandwidth). The system needs to be able to process the data as fast as the hosts want it. Systems are typically configured for the needs at and hopefully for some time after purchase. But systems do not have unlimited performance; depending on how long the system has been around and how many applications have been added that were not foreseen, there might be performance issues. And remember a NL drive typically has one-third to one-fourth the performance of a FC drive: How active are the various log volumes? While you can move the log volumes from FC to NL CPG, you might also get your phone to ringing as applications slow down because the log volumes are now quite a bit slower.

All of the following can be done using the SSMC browser interface. I use command lines since fundamentally the GUIs all send command lines to the system anyway. An SSH terminal interface (PuTTY on Windows) can be your 'friend'.

If your system has the Dynamic Optimization license, the volumes can be changed on-line. Otherwise you will have to shutdown the applications using a given volume while you make a corresponding NL volume then use whatever is needed to copy the data from the FC volume over. Next update the host(s) to use the NL volume in place of the FC volume. And finally restart the application(s).

And repeat for the next log volume.

You should already have a NL R6 CPG (possibly called "NL_r6"). Assuming it exists, you can move selected volumes to the NL CPG (several "tunevv usr_cpg NL_r6 <VV_name>" and "tunevv snp_cpg NL_r6 <VV_name>" commands).
You could make another CPG. The command "showcpg -sdg" will show all the parameters and options on all existing CPGs, then use the command "createcpg <desired_bunch_of_options> new_CPG_name" to make a CPG similar to the existing NL CPG.
Note that 3PAR Best Practice is to keep the number of CPGs to a minimum, so if you can use an already-existing CPG, so much the better.
All that being 'said', I would look at the performance of the various logs volumes over, say, the last week or so and start tuning the volume with the lightest IO requirements. As each completes, check that the application has not suddenly gotten significantly slower. If it's significant enough, the phone will be ringing anyway as users call to complain.
Bottom line: There comes a point when you may have to add physical storage to a system.

Your second question: "...[W]hy do I now have doubled VV..."? What command are you using to see doubled VVs?
I strongly suspect what you are seeing are not the volumes but the paths from the volumes to the VMs. Normal storage practice for availability is to have at least two paths from a given volume to the host. Assuming fiber channel, there would be two SAN fabrics ("A" and "B", "red" and "blue" or whatever), each an independent mirror of the other. So VV - A_fabric - host is one path, VV - B_fabric - host is a second path.
Depending on how the fabrics are configured (zoned), some volumes may have more than two paths.

There are best practices whitepapers available at the Hewlett Packard Enterprise Information Library (http://h17007.www1.hpe.com/us/en/storage/info-library/index.aspx).
Start with the "HPE 3PAR StoreServ Storage best practices guide".

[ConnieM]

Hi Smith

I assume you are correct on my 2nd question, i see something like this, it seem volume has two name ..

EXCH_XMS01_A1_FCr6_TPVV_7400 dojcdhdc-3par Base — Thin 1 350.00 85%

RAID 6 (FC) EXCH_XMS01_A2_FCr6_TPVV_7400 dojcdhdc-3par Base — Thin 800.00 84%

RAID 6 (FC) EXCH_XMS01_Arch3_NLr6_TPVV dojcdhdc-3par Base — Thin 1 680.00 100%

RAID 6 (NL) EXCH_XMS01_Lag1_NLr6_TPVV dojcdhdc-3par Base — Thin 1 350.00 78%

RAID 6 (NL) EXCH_XMS01_Lag2_NLr6_TPVV dojcdhdc-3par Base — Thin 950.00 90%

RAID 6 (NL)

[Sheldon Smith]

Oops, sorry. Without any example of what you were seeing, I thought you were seeing the same volume twice. Each of those appears to be a separate volume. Not sure what is going on. What command did you use to get that list?

You should contact 3PAR Support and ask them. You may need some consulting to take a close look at your environment and help you get it documented and under control.