Array Expansion

To add additional enclosure you dont require down time . This is NDU process .

Yes NDU process will reboot Sp one after the other.

To create customized documentation that contains step-by-step instructions for adding a hardware part to your storage system please follow the link.

http://powerlink.emc.com/km/live1/en_US/Offering_Technical/Technical_Documentation/cx3_cust_doc_upgrade.htm

What is NDU process? I know that EMC software is refereed to as NDU. is this NDU powerpath

Please elaborate.

So SPa will migrate all luns to SPb reboot SPa then vice versa?

Can you advise of any additional checks prior to installing the DAE

NDU refers to Non-Disruptive Upgrade  , for further instruction please refer to the link mentioned above for generating customised  document .Also refer to solution article emc36714.

So no extra software is required to manage the LUNs during the SP reboots ?

NDU = Non-Disruptive Upgrade

SP-B will be rebooted first. All LUNs belonging to SP-B will be trespassed to SP-A. Once SP-B is back on-line SP-A will be rebooted next. SP-A will be trespassed to SP-B. Once SP-A is back online verify that the LUN's whos default owner is SP-A are now owned by SP-A.

Where can I find article emc36714.

Correct. Just make sure that each host that is connected to the SAN has failover software (i.e. PowerPath) loaded and that each host can see paths to SP-A and SP-B and you'll be good to go.

On powerlink: http://powerlink.emc.com

Where can I check that we have powerpath.

Will I be able to locate it in Navisphere. Does it ship with a particular version of flare

Ernes, Using your experience could you email me a check list that you would use to carry out a DAE install?

Yes, use the USM for adding DAEs, it will help to make sure you don’t cause an outage by accidentally configuring the DAE with a duplicate ID# or other common mistakes.

There is no SP Reboot required for a DAE add on. The DAE is connected to the end of an existing backend bus, daisy chained to the last DAE on that bus, or directly to the SP if a bus has no DAEs already connected. Since you are not removing any existing DAEs, there will be no cables disconnected and as such, no outage.

Same goes for adding drives, you can add drives any time without downtime as long as there are free slots.

The key thing to do is: Check the disk/flare interoperability matrix on PowerLink to be sure that your FLARE version supports the drives you are planning to add. A FLARE code upgrade can also be done with USM, but it WILL require an NDU, which reboots each SP one at a time. In order to be sure your hosts do not lose connectivity during an NDU, each host must have a SAN path to BOTH SPs, and have path failover software configured. (EMC PowerPath, Windows MPIO, Linux Native-MPIO, VMWare NMP, etc.) If properly configured, the hosts will follow the LUN trespass between the SPs during the upgrade.

Again, the DAE install and drive adds themselves do not require an NDU, they are simply added to the system hot.

Richard J Anderson

PowerPath is licensed by host and installed within the host operating system. It is independent of the storage array.

Richard J Anderson

Swathi and Ernes are stating that this is a NDU with storage processor reboots

Richard is stating that there is NOT an NDU just a hot addition to the Array.

who is correct here?

you don't need to reboot SPs when adding disk shelves if you current FLARE supports drive model you are trying to add.

Message was edited by: dynamox

CX3 UltraScale Series Storage Systems Disk and FLARE OE Matrix   -- A13

Richard and dynamox are 100% correct. No need to reboot the SP's. I was just giving you a definition of NDU and what will happen during an NDU process. Adding a DAE will NOT cause NDU.