what did support have to say about that ? I am sure there is always a patch

Hi Dynamox, The incident is being reviewed by engineering. Hopefully they will come up with something.

admingirl,

I know that you mentioned the "serial port", but by chance was it the service LAN ports instead on each SP that the tech had simultaneously connected to?  The reason I ask is that for a DAE add, there is no reason to connect to the micro-DB9 connection on each SP, let alone to both simultaneously.  Also, very few have 2 of the proper cables (DB9 to micro-DB9) and to be honest, when I'm on-site very few know where the original is when I need it (which is why I carry one of my own for those few times when I need to pop into the serial port on the array).

Therefore, I am thinking that the tech may have connected to the service LAN ports which are standard Ethernet ports on each SP.

Also, do you have a VNX 5700 or 7500?  If so, the service LAN ports are internally switched and as a result you do NOT want to connect them to an external hub or switch.  If you own a VNX 5100, 5300, or 5500 these are not internally switch so we don't give you the same warning as we do for the larger arrays.

Just a thought...

DB9 cable taking the entire array down, taking about a single point of failure

I am assuming not the DB9 but the service lan RJ-45 connections.

I suspect that the tech connected to the service LAN ports on each SP simultaneously on a VNX 5700 and 7500 and connected them to an external hub or switch and then his/her laptop.  This is expressly forbidden.

However, if it was the micro-DB9 serial ports, then we'll need to have engineering provide feedback as I suspect it is unrelated.

Oh dear, if that is the case then it is expressly not allowed on a 5700 or 7500

(these arrays have an internal lan switch)

In essence when this happens it causes a network storm and can crash SP's

Not too sure of the kb article but I know its documented in the procedures

Gearoid

Hi Dynamox and Christopher,

, I believe he connected to the service LAN ports on each SP using a HUB. Once the array went orange, I had to scramble to bring production back up so I did not bother to ask him particulars at that time. We met with EMC yesterday and expect an RCA report based on the dump information.. We have a 5700.  Admingirl

admingirl,

Thank you for confirming our suspicions.  I have included the specific document related to the service lan ports.  This is included when generating any document from the procedure generator where that activity involves using the service LAN ports.

Note the following section:

[...]

CAUTION:    Never connect a switch or hub between the two Service LAN Ports on any VNX5700 or VNX7500 storage system. Connecting a switch or hub will cause the internal ports to go up and down as shown in the event logs. Use a single cable for a connection with a hub or a direct connection from your laptop to the Service LAN port. In the past, connecting a Hub to the management module has caused a network storm. If this problem occurs, promptly remove the Ethernet cable. Reseat the Management Module if the port appears not to function. The Management Modules are hot-swappable and may be removed and reseated without affecting the array’s operation. Refer to Primus emc199379.

[...]

Hi Gearoid,Can you point me to the procedure where this is documented? I ask because, I had posed this question to the tech and account TAM and they told me that the procedure used with the HUB is used all the time. If I have a document, then I can present this when we go over the RCA. Many thanks!

Im going to reply to you privately just so I can get some SR details (I prefer not to get these posted in public)

I believe I came across your case this evening in our daily DU/DL review, but I want to be sure

Regards

Gearoid