Michael wrote: > > You can transfer some data from a tmpfs and measure the speed. If it gets > anywhere near 4.8 Gbit/s (600 MB/s) its a SATA 3. The delay in the kernel > picking it up at boot may be related to its size, but I have no experience of > such large drives to be able to confirm this.
When I started reading your reply, I realized I had that drive in the older external enclosure connected to my main rig. I'm not sure if it is SATA 3 capable or not. I need to research the specs on that enclosure. It works fine for running self tests tho. Anyway, I connected it to my NAS box. On it, it reads at speeds that make me think it is SATA 3. It's around 250MB/sec with hdparm -t which is normal on most all my rigs. I have several drives of various ages but most are fairly new on the NAS box. At the moment I'm transferring data from one set of 4 drives to a set of 3 drives. The original set of 4 is a mix of spare drives I had laying around that is the backup of my large directory. I came up with a fancy command, for me at least, to get the drives size and speed. This is what I get. The first is the OS drive. The last one is the new 20TB drive. root@nas ~ # smartctl -i /dev/sda | egrep 'User Capacity|SATA Version' User Capacity: 500,107,862,016 bytes [500 GB] SATA Version is: SATA 3.3, 6.0 Gb/s (current: 3.0 Gb/s) root@nas ~ # smartctl -i /dev/sdb | egrep 'User Capacity|SATA Version' User Capacity: 10,000,831,348,736 bytes [10.0 TB] SATA Version is: SATA 3.2, 6.0 Gb/s (current: 6.0 Gb/s) root@nas ~ # smartctl -i /dev/sdc | egrep 'User Capacity|SATA Version' User Capacity: 14,000,519,643,136 bytes [14.0 TB] SATA Version is: SATA 3.3, 6.0 Gb/s (current: 6.0 Gb/s) root@nas ~ # smartctl -i /dev/sdd | egrep 'User Capacity|SATA Version' User Capacity: 8,001,563,222,016 bytes [8.00 TB] SATA Version is: SATA 3.1, 6.0 Gb/s (current: 6.0 Gb/s) root@nas ~ # smartctl -i /dev/sde | egrep 'User Capacity|SATA Version' User Capacity: 14,000,519,643,136 bytes [14.0 TB] SATA Version is: SATA 3.3, 6.0 Gb/s (current: 6.0 Gb/s) root@nas ~ # smartctl -i /dev/sdf | egrep 'User Capacity|SATA Version' User Capacity: 16,000,900,661,248 bytes [16.0 TB] SATA Version is: SATA 3.3, 6.0 Gb/s (current: 3.0 Gb/s) root@nas ~ # smartctl -i /dev/sdg | egrep 'User Capacity|SATA Version' User Capacity: 16,000,900,661,248 bytes [16.0 TB] SATA Version is: SATA 3.3, 6.0 Gb/s (current: 3.0 Gb/s) root@nas ~ # smartctl -i /dev/sdh | egrep 'User Capacity|SATA Version' User Capacity: 16,000,900,661,248 bytes [16.0 TB] SATA Version is: SATA 3.3, 6.0 Gb/s (current: 3.0 Gb/s) root@nas ~ # smartctl -i /dev/sdi | egrep 'User Capacity|SATA Version' User Capacity: 20,000,588,955,648 bytes [20.0 TB] SATA Version is: SATA 3.3, 6.0 Gb/s (current: 3.0 Gb/s) root@nas ~ # This is the PV's in order by the pv name. root@nas ~ # pvs -O vg_name PV VG Fmt Attr PSize PFree /dev/sdc1 backup-vg lvm2 a-- 12.73t 0 /dev/sdb1 backup-vg lvm2 a-- <9.10t 0 /dev/sde1 backup-vg lvm2 a-- 12.73t 0 /dev/sdd1 backup-vg lvm2 a-- <7.28t 0 /dev/sdf1 backup2 lvm2 a-- 14.55t 0 /dev/sdg1 backup2 lvm2 a-- 14.55t 0 /dev/sdh1 backup2 lvm2 a-- 14.55t 0 root@nas ~ # I'm surprised by the speed of the OS drive tho. It is a SSD drive. Anyway, as one can see, some drives are connected at SATA 2 and some at SATA3. Now I noticed, the 4 drive set is connected to a PCIe card. The three drive set and the OS drive is connected to the mobo itself. This makes me wonder, is the mobo ports only SATA 2? I went and looked at the manual that shows a block diagram and what is what. Sure enough, the mobo ports are SATA 2 or 3GBs/sec. Well, that explains that. It also explains a lot of other things. I thought encryption was slowing things down since the CPU doesn't have built in AES support. Turns out, the mobo SATA ports are slower than expected. I wanted to test two theories. I left the power connected to the newest 20TB drive that is usually slow to respond. I disconnected the data cable from the mobo port and plugged it into the PCIe card. Guess what, it had no slow to respond message and it connects at SATA 3 or 6GB/sec as it should. So, the new 20TB drive does take a little longer to power up since it connects fine when power is left applied when moving data cable. Also, the mobo is SATA 2 or 3GBs/sec. The PCIe card is SATA 3 or 6GBs/sec. That explains a LOT. It explains why my backup is slower than expected. It explains why drives are not connecting at the faster speed as well. It certainly explains why the SSD is connecting at SATA 2. I may as well go back to the old spinning rust drive since the mobo is the bottleneck on this. Oh, this is my little command to list speed of drive when the newest 20TB is connected to the PCIe card. root@nas ~ # smartctl -i /dev/sdj | egrep 'User Capacity|SATA Version' User Capacity: 20,000,588,955,648 bytes [20.0 TB] SATA Version is: SATA 3.3, 6.0 Gb/s (current: 6.0 Gb/s) root@nas ~ # So when connected to the PCIe card, drive speed is fine. What I learned from all this, I need to connect all my hard drives, data ones not the OS, to the PCIe card. I can also switch back to the old spinning rust drive for the OS. No wonder I didn't notice any increase in boot up time. I need to find a mobo/CPU/memory combo that supports SATA3 and has a few PCIe slots for a NAS box. This older system is getting a bit dated. Slower than I'd like too. I can't believe that mobo is SATA 2 tho. I could have sworn it was SATA 3. Dale :-) :-)
