I am not sure what happened to your card but it sound like the process misidentified it, resulting with bricking the card.
But don’t lose your hopes. i know that the folks from Mellanox support can bring back to life some of those cases.
Please open a support ticket with Mellanox support (email mailto://support@mellanox.com/ or web http://support.mellanox.com/ ) and somebody will give a hands.
I am afraid not yet. Since one IB HCA on the blade server is brick now, I think I should reflash the default firmware to the HCA somehow. However, I can not find a simple solution for this, such as a jumper on the HCA, OEM device by DAWNING, used for loading the HCA with the default firmware burnt on the HCA’s flash. There is no IB devices in the PCI list now. lspci | grep InfiniBand returns nothing.
I tried to install other operation systems on another node, Redhat 6 and 5. However, it failed to load the IB driver.
The firmware version 2.5.8, I do not know if I can use these IB HCA with OFED 1.5 or Redhat 6/5.
I despair of the upgrading program. One HCA lost is enough.
As an idea, do you have a normal PC or non-blade server around, with a free PCIe slot (PCIe x16 or PCIe x8)?
If you do, then it might be a better idea to update the firmware in your other cards with that instead of in the blade server.
I can give you the exact instructions for updating the firmware in a non-blade server (using either RHEL or CentOS versions 6.3 or 6.4). I have very similar cards here. Also MHGH28-XTC, but a different hardware revision (mine are MT_04A0120002, so different firmware needed).
After the firmware is updated, you could then see if the cards work properly in the blade server with Scientific Linux.
Btw, with the PSID for your cards, are you reading it from the sticker on the back of the card or did you get it from somewhere else? Just wondering if you might have gotten it from the wrong place, and therefore downloaded the wrong firmware. (unsure)