HP2910al + 10GB module j9008A >> reboot of switches

[sav azerty]

Hi,

My configuration:

one c-class c3000 with 2 virtual connect flex10

4 switches 2910al:

2 for lan

2 for iscsi

2 hp lefthand p4300 g2

3 bl465g7 with vsphere 5.1

all networking ar 1GB

Yesterday, i try to migrate the iscsi to 10gb

We bought 4 10gb modules j9008a. Lefhtand have already 10gb nics, each vc have 2 x-port unused

switches:

upgrading firmware to w15.08.0012

plug the 10gb module offline

restart ok, configure vlan, flow control

i disconnect the 2gb trunk between the 2 switches and connect one 10gb connexion (no trunk, link ok) with DAC 1m

no problem

lefhand p4300:

i upgraded the emulex 10gb firmware card: ok

connect 10gb nics with DAC 3m

i reconfigure a new bond (alb) with the 10gb nics: difficult, lost of ping, but finally it's ok

i did the same thing on the second lefhand: same difficult but finally it's ok

i thinked its the upgrade process to 1gb from 10gb that is long and difficult

virtual connect:

I connected 1 10gb DAC by virtual connect and configure ethernet network...

I need to reconfigure profiles of lames, and in the menu i saw "degraded" , i searched on discover it's iscsi is degraded (i have not remove this 1gb uplink)...

when I saw on switch, i saw orange/led, on the switch was rebooting

2mn after, the second switch was rebooting too..

On a physical server, who access iscsi network behind a router, i can ping switches, but  not lethand

On a notebook who is connect directly on a iscsi switches, i cannot ping switches, but lefthand ok !!

I decided to back to original configuration:

- lefthand to 1gb connection

- disconnect all 10gb DAC

- remove 10gb module

After this, communication between all servers, san and cie its ok, start lame servers, storage access ok

Do you know a major bug with this module j9008a and 2910al ?

[Sbrown]

Nope it is solid. It sounds like you need to RMA the defective part.

Make sure you are on the latest 15 series firmware (no java! web interface)

collect show tech all and open a ticket if you are having a reboot issue.

Had a bad 2920 do the same thing recently and it was just defective (switch itself).