The Synology RS3614xs+ has wrong HW specifications

Hi @ all.

the medium Business / Enterprise-level Synology RackStation RS3614xs+ has listed wrong tech specs on the synology website as well as almost anywhere else. This could be a slap in the face for some businesses, who bet on the new feature-rich and quite low-medium priced SAN/NAS solution.
But first let me explain…

This RS is on the market for about a year by the time of this post, and I could not find any information on the web, so I was posting in the synology forums here and here, but here as well.

First things first, my setup:

Synology Model Number: RS3614xs+ Firmware Version Number: irrelevant HDD Model Number: irrelevant Number of HDDs: irrelevant RAID Type: irrelevant Network Bandwidth/MTU: 2 cards withe 2x 10GbE each. Services enabled: irrelevant Hibernation time: irrelevant External Devices: 2x 10GbE 8-Port unmanaged Netgear switch

I have two of those listed devices here atm, for creating a 2-node cluster with Synology HA.

My 10GbE-Setup was planned as shown here:

 The connecting 
network goes here       (Internal Network)
      |    |
VM-Host1  VM-Host2      (Virtualisation Cluster)
      | \/ |
      | /\ |
 Switch1  Switch2      (Redundant Switches) 
      | \/ |
      | /\ |
  Syno1====Syno2        (Storage Cluster)

As you can see, I have a 10GbE HB bonded connection and one 10GbE planned to either switch for redundancy. I am talking about the green part only, to clear things up.

Got the boxes, set them up and faced multiple different issues.
I have had reduced net speeds on one of the four interfaces (shown with orange instead of greeen LEDs and the deactivation of either NIC in one of the two boxes, depending in which slot I put them.

Now, HCL is an issue for Vendors, I informed myself beforehand as you can imagine:

as listed here you can read:

Four 10GbE ports and SATA III 6Gb/s support
In addition to the default 4 Gigabit ports, RS3614xs+ comes with dual PCI Express Gen3 x8 slots, allowing installation of dual 10GbE Network Interface Cards (NIC) for maximized bandwidth.

Screen:

Synology Screenshot
Four 10GbE… – still wrong.

or here

Others
LAN Number (RJ45) Gigabit X 4 (supports 2 X optional dual-port 10GbE add-on card)

This was removed from the Synology website. – Not seriously though, they still say PCIE-8x – 2 (one with 8x the other with 4x) …you gotta be shitting me. The part in brackets only declines the main statement of how many PCIE-x8 are available, and there it still says: ‚2‘
Get this right, finally!

Screen:

Synology Screenshot
Edited by Synology

and here

Intel Ethernet Converged Network Adapter X540-T2 X540T2 2 X 10GbE 10GBASE-T RJ45 PCIe 2.1 5.0GT/s x8

which are the ones I purchased.

I opened several cases with my supplier for changing the NICs and even a case with synology for changing the whole box. The orange LED was gone, deactivation of one NIC persisted.

During the process I should download a logfile (.dat) to send to synology. Well this is what I did. I took a look inside for myself as well, (smart me ;) and I found several entries regarding my issues.

ixgbe 0000:01:00.1: eth5: Network adapter has been stopped because it has over heated. Restart the computer. If the problem persists, power off the system and replace the adapter

Starting with the overheating, which was the reason for the deactivation over time, which I already guessed before even looking at the log. The other entries regarding the NIC’s really have gotten my attention, though:

 [ 12.740763] ixgbe 0000:02:00.0: (PCI Express:5.0GT/s:Width x4) a0:36:9f:4f:30:44
 [ 12.924663] ixgbe 0000:02:00.0: eth6: MAC: 3, PHY: 3, PBA No: G45270-003
 [ 12.931356] ixgbe 0000:02:00.0: PCI Express bandwidth of 16GT/s available
 [ 12.938130] ixgbe 0000:02:00.0: (Speed:5.0GT/s, Width: x4, Encoding Loss:20%)
 [ 12.945251] ixgbe 0000:02:00.0: This is not sufficient for optimal performance of this card.
 [ 12.953669] ixgbe 0000:02:00.0: For optimal performance, at least 20GT/s of bandwidth is required.
 [ 12.962607] ixgbe 0000:02:00.0: A slot with more lanes and/or higher speed is suggested.

Now, what’s that supposed to mean – PCI-E Width x4 ? Shouldn’t it be x8 as stated above?
Looking a bit longer I found the other slot initializing the other card:

[Fri Oct 31 13:18:08 2014] ixgbe 0000:01:00.0: (PCI Express:5.0GT/s:Width x8) a0:36:9f:4f:30:44
[Fri Oct 31 13:18:08 2014] ixgbe 0000:01:00.0: eth4: MAC: 3, PHY: 3, PBA No: G45270-003
[Fri Oct 31 13:18:08 2014] ixgbe 0000:01:00.0: PCI Express bandwidth of 32GT/s available
[Fri Oct 31 13:18:08 2014] ixgbe 0000:01:00.0: (Speed:5.0GT/s, Width: x8, Encoding Loss:20%)
[Fri Oct 31 13:18:08 2014] ixgbe 0000:01:00.0: eth4: Enabled Features: RxQ: 8 TxQ: 8 FdirHash RSC
[Fri Oct 31 13:18:08 2014] ixgbe 0000:01:00.0: eth4: Intel(R) 10 Gigabit Network Connection

Well, that’s odd. The other slot seems correct.
Now I started to investigate, before I did not have a reason to question synology, but merely troubleshooting network connectivity.

Why would synology’s tech dept. write here on p. 12 one should add a PCI-E card in the white slot first for best performance, when they both are PCI-E x8?
Well, then, I opened the box and made a photo:
On the bottom left, you see the model number, the PCI-E specs on the right between the slots.

IMG_20141111_111110627_HDR_smOUCH! I don’t know wether the overheating comes from having too little bandwidth but this is just not what synology promises with advertising 2x PCI-E x8 and supporting 2x dual-port 10GbE.

I really am (was) a happy customer of synology products.
But unfortunately, reality hit me in the face with this one. Not meaning to be rude, it’s just disappointing.

Please synology, fix this issue asap. On your HCL, Website, with all suppliers and customers. My supplier, thomas-krenn already is in touch with european synology support, but this could mean a major recall program for all RS3614xs+ worldwide.

Regards,
erdnuesse

UPDATE 1: I added screenshots of the wrong parts, as far as still available. One of them got already updated, but the main one is still there. – Shoot! Come on, synology.