r720 with integrated S110 Integrated Controller

You can refer "Figure 3. Cabling Diagram—3.5 Inch (x4) SAS/SATA Backplane—PowerEdge R720" in below link for connecting back plane to software RAID. It is only supported on 3.5 X 4 backplane. For this to work you need to have cable wit part# MPXC8

https://www.dell.com/support/manuals/en-us/poweredge-r720/720720xdom/removing-the-hard-drive-backplane?guid=guid-460ea053-a7d7-4d92-aed5-faed86f44d0e&lang=en-us 

If you like to replace the controller I would recommend to replace with Mini Card only instead of PCIe adapter card. Below are part# on H710P mini card

TY8F9 - ASSY,CRD,CTL,H710P,1GB,NV,MCRD
TTVVV - ASSY,CRD,CTL,H710P,1GB,NV,MCRD

Hmm  okay so I have the 3.5x8 SAS backpane and it appears I currently have a 0MX3P7 going into the wider plug on the system board just like shown in Figure 2 of the link you sent.

I actually bought the TY8F9 and installed it but that's when I get the "PCI learning error". Does this mean I can only use the TTVVV model?

I'm trying to use all 8 drives instead of 4 if possible. Obviously if not possible then it is what it is I'll just use 4.

lol well that's kind of another issue I'm trying to fix by installing windows on the server to run the driver and BIOS updates since the LifeCycle controller errors out when downloading the updates from dell's FTP.

I'll have to buy the 3.5x4 mini sas cable to install an OS on the device to update the system I guess. Unless you know of a workaround for this RAID issue? (no drives populate when in AHCI or ATA mode)

Hello, thanks for choosing Dell. If I may, first it could be a hardware issue. As you can see from the image attached H720 isn't supported and part TY8F9 is H710, not H720. But another opinion is, if you see a PCI error, it also can be something else (riser, etc...) Please let us know if you have any further questions and have a good one.

Hi, if the disks are not detected with the S110 now, it is possible that H710 RAID won't make any difference, I'm afraid.  It also means the issue here isn't the controller, it could possibly be something else.

're not talking about the H720 whatsoever. Only H710 models. So I actually got a good deal on a PCI H710P with 2 mini sas plugs on it that I'm thinking of trying to run the 3.5x4 cables to the card.

The only issue I see here is the fact I have an integrated S110 RAID controller, is that going to completely take over for RAID or if I try and use this H710 PCI card will it take over in the BIOS?

Really at this point I'm just trying to make these drives usable with RAID however I can. I'm donating this server to a friend and he's going to need the local drive space.

Um okay I'm guessing you didn't read the replies from Shine K. The onboard RAID uses the 3.5x4 cables which I do not have, I have the 3.5x8 cables which are used with the H710 controller. The server did not have one when I obtained it so I bought one, it is authentic and used but it is the correct part. It throws a PCI Learning error.

So again I'll ask, can I used the PCI card H710P in RAID mode with this integrated controller and not have issues with the S110 taking priority over the H710P card. I can buy 2 3.5x4 SAS cables and run them directly into the H710P card, but I have no clue if I ca do that with an integrated controller on the motherboard and force it to use the add on card instead.

Also I know this is not a recommended way from Dell or what not but this is an out of warranty old used server so at this point it just needs to be usable in whatever way possible.

In regard to your firmware, if you have iDRAC access, you can upload the .exe update files there and not need to worry about connecting to the FTP server. If you don't have iDRAC access, there is also a bootable firmware image located on the support site that you could use. That image is essentially a bootable Linux image with the update repository for a specific model and an execution script. If you're interested, I can help you find the file.

From the way I'm reading your post, and forgive me if I've misunderstood, but it seems to me that you're looking to keep the current hardware and cabling configuration as-is, then activate the S PERC to control storage, since the H710P is having problems. If that's the case, then no, that won't work. The S110 is an embedded controller, so if storage routes through the H710P, it may as well not exist to the S110. If you have a controller with no storage running to it, you may as well remove it/deactivate it, depending on how you decide to proceed. The S110 can't act with the H710P as a middleman, so you may as well remove it. 

With the H710P throwing that error, it's going to do you no good in the server at all. It needs to either be replaced, or you'd need to downgrade to the S PERC and use the other cabling. If the goal was to make the server usable with it's current hardware set, there's nothing I could recommend. I'd happily share an anecdote from a time from a phone or forum support case,  but there's nothing I could provide at that level either. The only way I see to making the system usable will be to replace the controller, or to replace the cabling and use the S110. I'd certainly recommend the H710P far and away above the S110, but that's a different discussion. For a system that you know will work reliably, you'd really need to cable to the H710P and use that, or cable to the embedded controller and use it (and also set embedded SATA to AHCI to turn the S110 off).

If any of this is unclear, please feel welcome to follow up with me.

Okay awesome yes you kind of know what I'm trying to say/do. So I would MUCH rather keep this H710P and use it but are you saying the PCI learning error is indicating a hardware fault on the H710P itself or potentially a BIOS/driver issue?

I do have iDrac7 thankfully, but the exe just gets uploaded via the Update menu? I've never known I could do this with executables which is actually awesome if it works.

To my knowledge I cannot remove this integrated S110 correct? If I can I would absolutely LOVE to remove it and retest with this H710P (non PCI card model). I have the 3.5x8 cabling in this already so I'm hoping to stay with that and get the H710P issue fixed or yeah replaced if it is hardware failure but I'm not sure how to test that when I can't get into the BIOS.

Okay that is awesome, never nkew you could upload the BIOS update exe through iDrac and have it process the update.

So I have updated my BIOS, iDrac and OS Collector since that was the only drivers not up to date and I have installed the H710P back in. The exact error I am still getting is:

(Cut off on screen assuming it's ULD-LD shows) ULD Verison: 103
PCIe learning Error: Integrated RAID

This is after setting AHCI mod for ATA, turning it off, installing and then booting up to the above error.

In case of a hardware fault, I'd expect it's more likely the card than the board or slot. Still, flashing the BIOS, testing another PCI slot, and testing a different PCI card are all good troubleshooting steps to try to confirm. Speaking to the training error specifically, my understanding of that error is essentially that the system is trying to interface with the card, but not getting return data, or not getting the return dat it was expecting, at least. Flashing the BIOS would directly impact that process, so it can be a factor, I just wouldn't expect it to be the issue. I mostly wanted to touch on this because it'd been mentioned earlier in the thread.

Regarding the iDRAC7 and updates, yeah, it's pretty awesome as far as I' concerned. Being able to upload those .exes has been super helpful.

Finally, regarding the S110, correct. It can't be removed because it's part of the motherboard. You can turn it off by changing the Embedded SATA setting from RAID to AHCI (or ATA), though. 

Not being able to get into the BIOS is really the only red flag to me that the issue may not be the card. It may be a bit of a pain in the butt, but you might consider taking the server to a minimum to POST configuration, then trying to access the BIOS again. While you're taking it to that config, I'd perform a power drain and an NVRAM clear both just to be thorough. Once you're at the minimum to POST, power it back up and see if you can access the BIOS. If you still can't, then I'd suspect a board based issue. Plus, it sounds like this card was a recent order, so performing these steps may also help you in having your seller send a replacement.

Overall, the best troubleshooting recommendation I could make would be:

1) Take the server to the min to POST and attempt to access the BIOS. If you're able to get in, then the board shouldn't be having issues, but that's not a guarantee.

2) If you were able to get into the BIOS, I'd disable the S110 by changing Embedded SATA to AHCI (If you do the NVRAM Clear, this should reset the BIOS to default values, and thusly turn the S110 off)

3) Shut down again and reinstall the H710P expecting a training error. Provided you see it, shut down again and test in another slot. Provided you see the error again, I'd be heavily inclined to fault the card. Though testing another PCI device in these slots would also be ideal. Testing this H710P in another server would also be great, if possible.

Finally, if you'd like, I'd be happy to look over a SupportAssist collection from your server, if you'd like to provide one. The SupportAssist app will collect the hardware config, the system logs and the Lifecycle logs (it can also collect the RAID log from the PERC, but that won't work with a controller that isn't responding. Won't help us here, but it is useful to know). It can be exported from the iDRAC web interface. Send me a PM when/if you decide to export one. I'll be on shift for roughly the next 3 hours.

Let me know if I missed anything.

Just updating the thread:

I have talked with Dylan via PM and sent him logs and SupportAssist for potential diag. I have also ordered a new H710P to see if the one I have now is in fact dead/faulted.

Hi RumspringaHQ,

After looking over the files you shared, I'm really inclined to believe that the controller is at fault here. One red flag I noticed was that in the LCC log, the PERC is misidentified as the wrong device. If there's any PCI device around that you can replace this controller with just to see if you get a training error, it would be helpful, but at this point in time, I'm much more inclined to fault the controller. Without another PCI device, we can't really be certain, but my path forward would be to test a different PCI card, or to pursue having the seller send you a replacement card. The card is identifying as an generic HBA device and has been doing so since May 2nd. It doesn't see it as either a PERC, or as an LSI device. Considering you first posted on May 7 and it doesn't look like updates were attempted until May 12, that would also have me leaning toward this card arriving to you with issues.

The system is also complaining about the cabling, but I believe that's going to be related to some of the cable routing and troubleshooting from earlier. I don't think this is something we need to be worried about at this time.

I think your best next step is to install another PCI device if you can, just to prove slot health, but if one isn't available, go ahead and get a replacement card. If the system itself were having issues, I'd expect this to have problems with multiple different PCI devices or an inability to see devices at all. Instead, what we're seeing is that the system sees a generic HBA device and not a PERC. If the system is able to spin up something like an Intel NIC and say, "Yeah RumSpringaHQ, that's definitely an i350/X540/X710," then we'd know pretty much for sure. I obviously can't be sure at this time, but I am heavily inclined to believe the PERC is faulty.

Awesome! Well not awesome that that the controller is probably a piece of junk BUT I have a new one on the way!

This did at one point have an HBA installed in one of the PCI slots so that could be left over from that but as I said there's a new H710P on the way!!! Hope has been restored and I thank you Dylan and everyone else from Dell that has helped comment and trace the issue down to what we have now.

I do have a JJ8XD spare not doing anything but if I decide to use this I'll have to buy 2 of the 3.5x4 cables to be plugged into this card. Unless Dell has some voodoo magic like iDrac update where I can plug this in without SAS cablers connected to the card and it mysteriously works lol.