RN316 ESATA port speed

[Laserbait]

Hi there!   I was wondering what the eSATA port speed is on the RN3xx series is.   It doesn't state what speed it is in the hardware guide, so I'm guessing that it's SATA1 - 1.5Gbit, but I wanted to confirm.

I see this in my logs:

ata7.00: SATA link up 3.0 Gbps (SStatus 123 SControl 320)
ata7.01: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
ata7.02: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
ata7.03: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
ata7.04: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
ata7.05: SATA link up 1.5 Gbps (SStatus 113 SControl 320)
ata7.15: SATA link up 3.0 Gbps (SStatus 123 SControl 0)

ATA7.00-04 I believe are my disks (HGST HUH721010ALE601) in the EDA-500.
ATA7.05 and ATA7.15 I'm sure are the eSATA port and the other possibly maybe a SES (SCSI Enclosure Services) chip.  I just don't know which is which.

[Sandshark · Accepted solution]

I believe the eSATA speed is SATA2 (3Gb) -- it is on a 516.  But with an EDA500, it's multiplexed out to 5 drives, so the effective speed is far less.

[Sandshark · Accepted solution]

I believe the eSATA speed is SATA2 (3Gb) -- it is on a 516.  But with an EDA500, it's multiplexed out to 5 drives, so the effective speed is far less.

[Laserbait]

Thanks SS.   I was hoping that it was.  I'm currently going through a disk rebuild, and it's doing about 25MB/s.  There are a bunch of other things using the disks too (like ReadyDR backing up the data on there), so I'm not entirely surprised by the rebuild speed.

/home/admin# cat /proc/mdstat
Personalities : [raid0] [raid1] [raid10] [raid6] [raid5] [raid4]
md1 : active raid10 sdf2[5] sde2[4] sdd2[3] sdc2[2] sdb2[1] sda2[0]
      1566720 blocks super 1.2 512K chunks 2 near-copies [6/6] [UUUUUU]

md126 : active raid5 sdl3[7] sdg3[6] sdk3[8] sdj3[9] sdi3[5]
      39046348288 blocks super 1.2 level 5, 64k chunk, algorithm 2 [5/4] [U_UUU]
      [=====>...............]  recovery = 29.4% (2873515392/9761587072) finish=5058.6min speed=22693K/sec

md127 : active raid6 sda3[0] sdf3[5] sde3[4] sdd3[3] sdc3[2] sdb3[1]
      31236703232 blocks super 1.2 level 6, 64k chunk, algorithm 2 [6/6] [UUUUUU]

md0 : active raid1 sdl1[20](W) sda1[14] sdg1[19](W) sdi1[13](W) sdj1[10](W) sdk1[11](W) sdf1[17] sde1[16] sdd1[12] sdc1[15] sdb1[18]
      4190208 blocks super 1.2 [11/11] [UUUUUUUUUUU]

unused devices: <none>

And the CPU is pretty backed up, but does still have some idle cycles.

/home/admin# top
top - 09:57:02 up 17 days,  4:06,  1 user,  load average: 8.52, 8.95, 8.79
Tasks: 276 total,   3 running, 273 sleeping,   0 stopped,   0 zombie
%Cpu(s): 12.2 us, 42.8 sy,  0.0 ni, 24.3 id, 15.7 wa,  0.0 hi,  5.1 si,  0.0 st
KiB Mem:   2032656 total,  1947168 used,    85488 free,     1220 buffers
KiB Swap:  1566716 total,       72 used,  1566644 free.  1370356 cached Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 8287 root      20   0   50276   4228   2028 R  84.0  0.2 613:48.09 ssh
 8291 root      20   0   24000   1304   1124 R  69.9  0.1 499:33.88 btrfs
 1640 root      20   0       0      0      0 S  24.4  0.0   1105:25 md126_raid5
 8281 root      20   0   57956   3752   2068 S  21.4  0.2 148:57.96 replisync
 1933 root      20   0       0      0      0 S   7.9  0.0   3:44.07 kworker/u8:0
  668 root      20   0       0      0      0 S   5.9  0.0   1251:26 kswapd0
 9573 root      20   0       0      0      0 S   5.6  0.0   0:39.48 kworker/u8:3
17173 root      39  19       0      0      0 D   5.6  0.0 135:24.56 md126_resync
10668 root      20   0       0      0      0 S   5.3  0.0   0:14.08 kworker/u8:1
11016 root      20   0   28932   3276   2568 R   1.0  0.2   0:00.08 top
10334 root      20   0       0      0      0 S   0.7  0.0   0:30.20 kworker/u8:6
10990 admin     20   0  156120   7248   4196 S   0.7  0.4   0:00.02 apache2
    3 root      20   0       0      0      0 S   0.3  0.0  50:15.47 ksoftirqd/0
   13 root      20   0       0      0      0 S   0.3  0.0  49:59.46 ksoftirqd/1

I wonder if upgrading memory from 2GB to 4GB would have much of an impact on operations like this (in general to RAID rebuilds). I know the eSATA link is probably the bigger bottleneck, but I can't really do much about that currently.

[Sandshark]

Sync and scrubs on an EDA500 are abysmally slow due to the eSATA multiplier architecture.  I ultimately retired both my 516 and EDA500 in favor of a 12-bay rack-mount unit (you can just move all the drives), but that's not an option for everyone, given the size and fan noise.