08-02-2019, 12:04 AM
(This post was last modified: 08-02-2019, 12:12 AM by stuartiannaylor.)
(08-01-2019, 08:32 AM)ZeblodS Wrote: @stuartiannaylor: Just to answer a few of your talking points.
I was thinking of using a regular ATX power supply for several reasons, the main one is that it's cheaper to get 12V and 5V at pretty high amperage.
For instance I can find an brand new EVGA 600W with a single 12V49A rail and a 5V20A rail for less than 50€. More than enough to power safely the ROCKPro64 and 10 to 12 hard-drives.
Plus it's rated 80+, meaning less electricity wasted. And usually brand names power supply don't destroy all your hardware when they fail (which all power supply end up doing eventually, even brand names one as I have already experienced it after 7 years 24/7 powered on).
It has 6 SATA power connector, that can be doubled using adapters like these.
Power supply, for machines running 24/7, is really not something you can really spare expenses on.
I looked over for a USB3.0 solution, using a PCI-E like the one you linked. UASP adapter with 12V input for 3.5" harddrives are not that cheap either, I have found models like this one for instance. Plus I'd need a few USB3.0 hub to connect 2 to 3 UASP adapters per USB ports on the controller, with maybe 5V external connection to not overpower the PCI-E controller with that much disks.
In the end it won't be cheaper than the IOCrest SI-PEX40097 SATA controller, and it's a less clean solution, but it's a fallback solution for sure.
For the casing, I'd still go the individual hot-swap from the front route as it's way more practical IRL, even if it cost more. The Orico one can be found for 8.75€ a pop when bought by 10 or more.
Software wise, I'd need to do some more research into it, and some testing before going toward a different solution than what I already use. But on first glance Snapraid seem to have many advantages indeed.
EDIT: the USB solution could be cheaper using something like these:
https://m.fr.aliexpress.com/item/32897789991.html
https://m.fr.aliexpress.com/item/32829472708.html
If you do decide to go for https://www.aliexpress.com/item/32546377052.html it will be interesting what benchmarks you get.
Highest I have seen have been a single Samsung NVME 970 SSD
Code:
Command line used: iozone -e -I -a -s 100M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2
Output is in kBytes/sec
Time Resolution = 0.000001 seconds.
Processor cache size set to 1024 kBytes.
Processor cache line size set to 32 bytes.
File stride size set to 17 * record size.
random random bkwd record stride
kB reclen write rewrite read reread read write read rewrite read fwrite frewrite fread freread
102400 4 52099 77524 104207 105324 48114 78504
102400 16 160544 231475 276300 278531 160180 233806
102400 512 738881 809485 685261 707192 676594 811610
102400 1024 795915 841642 708547 730494 693498 851120
102400 16384 1111519 1167200 1098700 1139913 1110364 1193693
iozone test complete.
From a while back when I tested my Marvell 88SE9235 4 port with just a RAID0 test I got with a test of 4x budget Integral P5 120gb sata SSD
Code:
Command line used: iozone -e -I -a -s 100M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2
Output is in kBytes/sec
Time Resolution = 0.000001 seconds.
Processor cache size set to 1024 kBytes.
Processor cache line size set to 32 bytes.
File stride size set to 17 * record size.
random random bkwd record stride
kB reclen write rewrite read reread read write read rewrite read fwrite frewrite fread freread
102400 4 33519 47927 52701 51023 26700 46382
102400 16 105763 132604 138080 155514 87026 135111
102400 512 276220 320320 311343 294629 267624 335363
102400 1024 493565 522038 463105 470833 398584 522560
102400 16384 687516 701200 625733 623531 555318 681535