Dell EMC's PowerMax migration: Let's just swaaap out this jet engine mid-flight

This month Dell EMC introduced a new technology and rebranded VMAX array called PowerMax. It's more than a software upgrade, though, meaning that you may need to replace a VMAX with a PowerMax array. Those who choose to upgrade have to swap out their old storage hardware due to an array architecture change – as occurred with …

[stiine]

seriously? and they charge for this?

Hmmm, sounds just like vMotion. I wonder why?

[Nate Amsden]

Re: seriously? and they charge for this?

Because many workloads don't run in a hypervisor

[Androgynous Cow Herd]

Technology marches forward

You can now allegedly do a Symmetrix upgrade the way Nimble storage has done generational upgrades for 5 years!

Personally, I would still call an downtime for the upgrade if I possible.

[Nate Amsden]

Re: Technology marches forward

Customers have been able to migrate from vmax to 3par non disruptively for years as well.

I am very curious to know the architectural change that requires data migration.

[vogie563]

This seems very similar to the PowerPath Migration tool included in their Multi-Path utility, just array based not host based. The host based tool is great for bare metal migrations. I wonder if this does the extra step of virtualization of SAN WWN's to avoid mass zoning updates?

[Anonymous Coward]

Non-disruptive data migration requires the target device to have the same WWNN and SCSI characteristics as the source device. From the host point of view, all you are doing is adding paths for the volume. Assuming you are using ALUA (and you most likely will be) the storage will tell the host which paths to use and the storage will handle the synchronisation of the data between the systems as it is copied across.

The only way this can be non-disruptive is for the new array to be able to pretend to be the old one in this manner. So if you're migrating from another vendor's system, you will need an outage at some point, to change the volume from being old vendor's volumes to being new vendor's volume (i,e, different WWNN and SCSI IDs).

Even if you can do it completely online, as Nate mentions, you should always schedule a controlled reboot anyway. It's better to do the reboot to check that everything works (and to document it) than to find out six months down the line that some of your hosts won't boot.

[IGnatius T Foobar]

Wow?

Congratulations, EMC (or should we call them Storagezilla?) ... you've invented a feature NetApp has had since they introduced Clustered OnTAP 8.x some number of years ago.

[smoked_2na]

The sky is not falling chicken little!

It's more than a software upgrade, though, meaning that you may need to replace a VMAX with a PowerMax array. <---- REALLY

This is like saying that Ford came out with a new car without a gasoline engine so you have to buy and move all your stuff to the new car today.

The synopsis of the story should have been, NVMe in the new re-branded DELL storage product, migration is covered.

[Anonymous Coward]

When You Bet Your Career on Symmetrix

I am sure PowerMax will do fine as many people paid a lot for the previous generations and to admit it was not needed could be career inhibiting.

PowerMax, isn't that a protein shake, and before EMC had ViPR it was a piece of exercise equipment. It would appear EMC marketing either don't have access to google or are lazy roid heads.

[Anonymous Coward]

well it's about time

but, don't think that's as simple as a Pure Storage NDU.

Just sayin!

[Anonymous Coward]

Re: well it's about time

Pure's NDU is data-in-place. No movement of data between arrays, because the controllers are upgraded one at a time in the existing array. I still find it hard to believe big boys like EMC etc are still pulling the wool over customers eyes.

[flyguy959]

Re: well it's about time

Data in place for //m series controller upgrades or to //x(which is basically a controller upgrade). Going from the old Dell based controllers on the FA-4xx series to the //m appliances was a NDM across arrays. You had to install HBAs into a reserved PCIe Slot on the new array controllers to handle the cross array data movements. There's a video series on taking a FA-4xx to a //m20 to a //m50 somewhere on youtube.

[SuaSponte375RGR]

Make no mistake, this is just lipstick on the same old EMC pig we all know. This is still the same old Dell EMC forklift upgrade. The customer has to still forklift upgrade due to HW changes and non-compatible hardware between platforms. The architecture is old archaic and marketing is lazy by just changing a component or two and rebranding again. Until they truely redesign both the hardware and software from scratch, it will always have limitations that require forklift upgrades.

Bottom line: swapping the entire system, that’s a forklift upgrade. When you can swap components and not have to perform a migration or get new serial numbers, that’s a true NDU platform.

Don’t be fooled. Pure Storage has true NDU and hardware protection for customers.