This article summarized the Chinese Ask the Expert event of VPLEX HA & Mobility platform. The content summarizes 5 Q&A which indicates the VPLEX implementation practice & recommendation.

Q1: Regarding the EMC best practice? It indicates VPLEX Frontend & Backend 00, 03 connect to one fabric, 01,02 connect to another fabric. In case some of customer has ports limitation of SAN switch, which causes VPLEX in “half configured”. I heard about the FE&BE 00, 01 are controlled by one CPU core, 02, 03 are controlled by another CPU, so the recommendation is just use 00 and 02 ports. Does it recommended by best practice?

A1:  In VPLEX architecture, FE&BE not only share CPU resource, so EMC doesn’t require any pair of ports have to be connected in same fabric.  EMC best practice recommends connecting FC00 and FC02 to same fabric, FC01 and FC03 connecting to same fabric. In your case of “half configured”, you can connect FC00 to one fabric and FC01 to another.

Following are the cabling diagram of 1, 2, 4 engines best practice.

1 engines:

2 engines:

3 engines:

4 engines:

Q2: During VPLEX implementation, how to ensure the business continuity? If the storage array attached to host through SAN which has not been added into VPLEX, how to ensure the data continuity during adding existing array under VPLEX? I mean the live data migration and down time consideration.

A2: Regarding the question about ensure business continuity during VPLEX implementation, the answer is” It depends”. If it’s initial setup, VPLEX does not contain any production data. If VPLEX is going live and new backend drive is added, I believe you can just follow the steps in SolveDesktop to make sure business continuity. Following are the VPLEX best practice & installation guide of VPLEX.

https://support.emc.com/docu9765_Technical-Notes:-Implementation-and-Planning-Best-Practices-for-EMC-VPLEX.pdf?language=en_US

https://support.emc.com/docu34005_VPLEX-Configuration-Guide.pdf?language=en_US

For the question about adding volume from existing array you ask, I’d like to recommend below steps.

Copy data to a new volume (same array)

Present new volume to VPLEX

Attach encapsulated VPLEX volume to host.

If you directly attach the original volume to VPLEX which will cause reservation conflict error because of host has place lock on the original volume. VPLEX can’t access it directly.  If adding VPLEX into an existing environment which host is writing data to volume, the down time is required. To minimize the down time, you need to consider the host cluster solution.

Q3: I have real implementation case which requires four 78GB drives. EMC recommends use two array as backend. But due to some limitations, my customer has to create four 78GB drives within one array then present to VPLEX. After VPLEX working well, then mirroring VPLEX Meta Volume to another leg or delete and create new Meta Volume. My question is if there is any risk to do this. Do you have any configuration process or method?

A3: Regarding the live data migration you ask, there are two methods to achieve it.

1.     Present a LUN from new array to VPLEX and attach to VPLEX as a Meta Volume Leg. After the sync complete, detach original leg. The Meta Volume is cross two arrays. But you should make sure the leg of original array is healthy during data sync. Otherwise, single point failure will cause systematic problem.

2.     Create a new Meta Volume legging to both array, then issue command “meta-volume move”, which will move in-memory data to new Meta Volume and activate the Meta Volume. Once the process complete, you can destroy old Meta Volume.

Q4: Are there any recommendation about SAN switch ports configuration of VPLEX?

A4: There is no any specific guide of SAN ports configuration about VPLEX. But I’d like to recommend you follow the VPLEX best practice guide.

https://support.emc.com/docu9765_Technical-Notes:-Implementation-and-Planning-Best-Practices-for-EMC-VPLEX.pdf?language=en_US

Q5: I have a VPLEX Local, 2 DS 300B and 2 VNX2 5200 arrays in my environment. If I try to copy files (ISO file) from host, then I disconnect one of VNX array during copy process. Any problem what if I run MD5 checksum between source and destination on the ISO file? My second question, how do I monitor my VPLEX environment? Is it supported email notification? What kinds of notification can I receive through monitoring?

A5: For your first question, if the Top Level Device which you copy ISO file is present by VPLEX and your VPLEX to VNXs connection is fully redundant. The disconnecting the one of VNX from VPLEX will not impact the MD4 checksum result. About your second question, EMC recommend customers to deploy ESRS gateway which will help you receive email notification first time any event occurs in VPLEX environment. You can also deploy Watch4New solution from EMC to have overall monitoring. 

Author: Fenglin Li

iEMC APJ

Please click here for for all contents shared by us.