June 29, 20224 yr I just recently bought a new 18 TB HDD to be added to my Unraid setup. As I had some older 4TB HDD laying around I added them as well. All new disks are connected using a PCI-E Sata card (BEYIMEI PCI-E X1 SATA - ASM1166+JMicron JMB5xx-Chip). During the pre-clearing of all drives in parallel Unraid reached roughly 70MB/s per drive. After the smaller drives have been pre-cleared the 18TB peaked at 270MB/s. I am now wondering where this difference of speed originates. Is this due to the age of the harddrives, the PCI-E controller, or is this just "normal"? If the PCI-E controller is a bottleneck, it would be easy to replace. If the harddrives itself are the bottleneck - well so be it. Can someone help me understand the differences in speed? Other hardware: Gigabyte Technology Co., Ltd. B550 AORUS PRO V2 AMD Ryzen 7 PRO 5750G
June 29, 20224 yr 1 hour ago, MrGumba said: ASM1166+JMicron JMB5xx-Chip This suggests the controller uses one or more port multipliers, if yes bandwidth of any SATA port connected to one will be divide by all 5 ports, diags would confirm.
June 29, 20224 yr Author Thank you Jorge. I just realisied that the 18TB is not connected via the card, but directly to the MB. So it seems that the card itself is a bottleneck, although it should have enough wiggly room with 6GBs to easily max out all of the harddrives currently attached to it. As the 4TB and 3TB HDD are only spinning at5400 RPM, I think it might be a combination of not so young HDD´s and the need to move data across the bus from the motherboard to the card. Lets see how much of a day-to-day impact this will have. If the problem becomes a daily issue, I will likely switch the SATA controller card to one that has been recommended in the forums. If this happens, I will report back with my experiences. [0:0:0:0] disk SanDisk Cruzer Fit 1.27 /dev/sda [1:0:0:0] disk ATA ST18000NM000J-2T SN01 /dev/sdb [2:0:0:0] disk ATA SanDisk SDSSDH3 20RL /dev/sdc [3:0:0:0] disk ATA ST18000NM000J-2T SN01 /dev/sdd [4:0:0:0] disk ATA ST18000NM000J-2T SN01 /dev/sde [5:0:0:0] disk ATA ST12000VN0008-2P SC61 /dev/sdf [6:0:0:0] disk ATA ST12000VN0008-2P SC61 /dev/sdg [12:0:0:0] disk ATA WDC WD40EFRX-68N 0A82 /dev/sdh [36:0:0:0] disk ATA WDC WD30EFRX-68E 0A80 /dev/sdi [37:0:0:0] disk ATA WDC WD40EFRX-68N 0A82 /dev/sdj
June 29, 20224 yr Yes, also note that the clear thread runs for all the disks at the same position, i.e., speed will be limited by the slowest disk at that point, screenshot appears to show the end sectors of the 4TB disks, since the 3TB dis already done, so 70/80MB/s at the inner sectors for those disks is normal, probably not the controller being the bottleneck in this case, still controllers with port multipliers are not recommended for performance and stability reasons.
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.