arm-push

Yes. I don't use a heat sink.

If anyone is interested in lowering the heat from the board, I've used a 5.9V/3.8A switchmode power supply with great success.
A 5.2V/2A PSU should be enough, though - my PSU just happened to be 3.8A.
The Voltage range for the PSU should be above 5.2V and below 12.5V. There is an on-board 13V zener, this is in order to try and protect any attached harddisk from a spike.
As for current, anything above 2A should be fine. If you're planning on powering a 3.5 inch harddisk from the board, stick with a 12.2V PSU.
-If you're planning on using a 2.5 inch harddisk, you can run on low input voltage between 5.2V and 12.5V.
 
Personally, I'll be using 2.5 inch WD Red drives because they consume very little power.
I might also add a liquid cooler (cheap aluminum blocks from eBay and a silent <5dB, <6W, 0 .. 8L/min pump) - I hate noise.
Note: I have not had any crashes or other issues with my 2GB boards - not even while running the boards on a 12.25V PSU with a 3.5 inch harddisk; the cooling is likely not even necessary, as the boards will be installed in an always-cold room. If, however, I add the cooling, I'll run the pump slowly, so it'll likely use less than 1W anyway.

I can confirm that the 4 Port SATA III 6G Mini PCI Express Marvel 88SE9215 Controller Card is working out of the box with EspressoBin. It is the same as this one known to work and tested by Igor (results for EspressoBin).
 
It is possible to use the mPCIe card inside the enclosure and loop through the 4 SATA cables to an external backplane - but then the internal 2.5" drive has to be removed from the enclosure of the EspressoBin due to space constraints. In that case it still can be used externally as a 5th drive if needed.

esspressobin is affected by insanely high delivery cost instead.

EspressoBin has Cortex A53 processors and is therefore not affected by Meltdown/Spectre attacks: see ARM link.

And guess what: I have a huge box here labeled 'PC-Schraddel' (PC junk), just checked it for those cables and to my surprise I found in there my EspressoBin and also the right cable with 2 female Molex jacks:

We have produced a few minimalistic housings for the EspressoBin (120x42x81mm) with space for a 2.5" drive using a CNC molding cutter.
Passive cooling of the processor and topaz switch by thermal coupling to the housing works well (housing temperature 31 degrees, ambient temperature 20 degrees).




 
 

We use systemd-networkd at Espressobin. Check those manuals/examples:
 
http://xmodulo.com/switch-from-networkmanager-to-systemd-networkd.html
https://wiki.archlinux.org/index.php/systemd-networkd
 
Leave /etc/network/interfaces empty.

I've compared the kernel config for mainline and Marvell's 4.4 and one thing that stands out in regards to the SD card issue is:
CONFIG_MMC_SDHCI_XENON
In the linux-mvebu64-default.config it is set to y, in linux-mvebu64-next.config is it not set
 
That could explain why it works to @umiddelb, his defconfig has it enabled too. I'll try it myself when I get the chance but that could be a few days away.
 
Edit:
I've just tried it on next and it works, just need to update fdt_name env var to point to the correct DTB
# uname -a Linux ebin 4.14.0-mvebu64 #4 SMP PREEMPT Mon Nov 20 08:02:32 UTC 2017 aarch64 GNU/Linux # lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT sda 8:0 1 1.4T 0 disk └─sda1 8:1 1 1.4T 0 part mtdblock0 31:0 0 4M 0 disk mmcblk0 179:0 0 29.8G 0 disk └─mmcblk0p1 179:1 0 29.5G 0 part /  

I'm using 1200_750 and I've never had any issues. I think it's the stable one.
With the settings I mentioned above, I can get around 78 to 80 MiB/s write rate over 1G Ethernet network.
As @tkaiser mentioned, I have done quick tests and it seems that the CPU is not powerful enough to exceed about 80 MiB/s. CPU utilization reaches 100%, mainly the application ( vsftpd in my case ) and kernel threads related to virtual memory contribute to that.