1

I have a new to me server I am trying to break in. One of the drive bays does not detect any drive. There are 24 bays total and 16 of them currently detect drives and allow setup of array, I have not yet tested the remaining 7.

I have the steps I am trying: 1. ticket with the vendor who just sold it to me. 2. try known working drives in the slot. 3. firmware updates. 4. hypervisor updates.

Anyone else have suggestions of what they do in this instance? Honestly I will probably just bypass using that slot if that is the only thing wrong here, but want to know if there are any further steps I can take in the field short of replacing the controller interface backplane board.

4 Spice ups
2

Are all 24 ports live, you’ve confirmed by removing the lid that there are enough SAS cables going to the front bays?

What controller is it, is it HPE Smart Array P408i-a?

This card supports 24 bays, but other cards may not.

Does the drive bay light up when it has a drive, but just isn’t seen by the controller, or does it not power at all? The latter being those ports may not be live or a cable may have come loose.

1 Spice up
3

yes. with cache and battery. I insert the drive and it lights up like all the others. it just doesn’t appear as available unless I connect it to another position. all 7 of the other connections on this backplane the drives are fine. I even looked to make sure there was no obstruction covering the connector.

4

The drive is confirmed good, known and working?

The same drive works in another slot or another drive also doesn’t work in 1 slot?

Any errors or clues in the iLO logs?

Is this bay for some reason disabled in BIOS or marked as a spare in the configuration - I assumed you reset/wiped the controllers config on receipt, but I could be wrong, you may have went with it out of the box.

Check the controller logs or confirm it isn’t marked as bad/spare/faulty, change the view in the drop down see if it shows under another view.

5

I honestly don’t know how to wipe out the configuration. That is probably a good point. I never checked to see if it was expecting to have some array in there or not.

6

In the ACU, there is a clear config when you select the controller. Typically on the right side of the screen. This will remove all RAID from the system, you’d have to then create a new array. Don’t do this is your system is in production, it erases the configuration so you have however many drives no longer in RAID/ You then setup your array with the number of drives you want and RAID type, but you should see the drives completely if 1 slot is assigned hot-spare. (You should see it anyway, but perhaps you’re missing it0.

1 Spice up
7

ok. sounds like a good place to start. thank you!

8

I think we have a winner. After re-creating the arrays a second time I was finally able to see the clear controller configuration option which I selected. After re-creating them a third time I showed one random drive failure somewhere, but after re-install of ESXi and reboot, this time I am seeing a drive installed in that bay using ssacli on the command line.

1 Spice up
9

Glad to see it’s solved.

10

alas, that was not the end of the story. I am day 3 of trying various hardware switcharoos with a similar server that is working with no errors.

here is what I have tried:

switched the drive cage and backplane out with one provided via RMA by the vendor
switched the SAS cables out between that bay and the SAS expander card
switched the SAS cables out between the SAS expander card and the controller board
switched the SAS expander cards

have not been able to migrate the error to the other server with any of this, so assume that the SAS cables work fine as well as the expander cards.

I am going to try now swapping the drive bays, then if that doesn’t work, swapping the smart array controller boards themselves.

anything else this could be that I might be missing?

Thanks.

11

I did this backwards. I swapped the smart array controller boards first. no luck, drive detection problem stays with original server.

I then swapped known working drive bays. Still detection problem stays with original server.

Now I have since I started this gone as far as taken working components from known working servers, moved them over and traded them for suspected not working parts in this server, and each time found that the original component works fine in another server and the working part from another server still gives the non detection in port 15 of the 24 port array.

Since I can’t get the problem to migrate to another server to ID a non working component, and since I still have the exact problem with every part in the chain swapped in the original server, could iLO or the firmware itself be “remembering” the problem? Any steps on how to do a complete factory reset of the whole server so I can start over fresh and see what happens?

Thanks.