Testing a SCSI SCA SSD from Red Rock Tech...

ElCoyote_

New member
May 25, 2022
8
9
3
Hi All,

This is probably going to be part of a larger post as the situation is still developping.

I'm attempting to replace the 20yo boot HDD in my Tezro with an SCA SSD from Red Rock.
I got one of their drives a few days ago (it was very expensive, think the current street price of an acard ARS-2160 or ARS-2320).

I bought the 960GB TLC SCA LVD version of this drive:

- The good:
Installation was very easy (just pop the drive into an Sgi caddy) and the drive is quite fast.
I used the clone_drive_any script from gijoe77 to clone the HDD onto the SSD, then pulled the HDD out and placed the SSD in slot #1.

Here are the diskperf numbers from that HDD:

Code:
[root@istar /d]# diskperf -D -W -n "SGI ST373405LC" -r4k /d/test
#---------------------------------------------------------
# Disk Performance Test Results Generated By Diskperf V1.2
#
# Test name     : SGI ST373405LC
# Test date     : Sun Nov 13 08:54:20 2022
# Test machine  : IRIX64 istar 6.5 07202013 IP35
# Test type     : XFS data subvolume
# Test path     : /d/test
# Request sizes : min=4096 max=4194304
# Parameters    : direct=1 time=10 scale=1.000 delay=0.000
# XFS file size : 1048576000 bytes
#---------------------------------------------------------
# req_size  fwd_wt  fwd_rd  bwd_wt  bwd_rd  rnd_wt  rnd_rd
#  (bytes)  (MB/s)  (MB/s)  (MB/s)  (MB/s)  (MB/s)  (MB/s)
#---------------------------------------------------------
       4096    0.68   24.14    0.69   10.63    0.77    0.85
       8192    1.34   41.86    1.40   15.25    1.49    1.69
      16384    2.61   55.96    2.86   20.21    2.93    3.23
      32768    4.99   55.86    6.03   23.73    5.52    6.21
      65536    9.16   55.95   13.52   25.82   10.00   11.12
     131072   15.76   55.96   17.73   33.44   17.12   18.39
     262144   24.57   55.95   35.74   35.75   26.08   26.91
     524288   34.13   55.96   35.81   35.78   35.08   36.09
    1048576   42.37   55.93   45.24   45.09   42.87   40.09
    2097152   48.08   55.94   52.06   50.99   48.19   47.95
    4194304   51.76   55.94   52.09   52.10   51.25   50.33
Here are the diskperf numbers from the Red Rock Tech SSD:
Code:
[root@istar ~]# diskperf -D -W -n "Red Rock SCA + Intel DS3 S4510 960gb" -r4k /d/test
#---------------------------------------------------------
# Disk Performance Test Results Generated By Diskperf V1.2
#
# Test name     : Red Rock SCA + Intel DS3 S4510 960gb
# Test date     : Sun Nov 13 08:15:21 2022
# Test machine  : IRIX64 istar 6.5 07202013 IP35
# Test type     : XFS data subvolume
# Test path     : /d/test
# Request sizes : min=4096 max=4194304
# Parameters    : direct=1 time=10 scale=1.000 delay=0.000
# XFS file size : 703856640 bytes
#---------------------------------------------------------
# req_size  fwd_wt  fwd_rd  bwd_wt  bwd_rd  rnd_wt  rnd_rd
#  (bytes)  (MB/s)  (MB/s)  (MB/s)  (MB/s)  (MB/s)  (MB/s)
#---------------------------------------------------------
       4096   13.00   14.16   13.04   12.20   13.30   11.30
       8192   22.65   23.99   22.64   20.82   22.47   17.67
      16384   34.49   36.52   34.64   26.70   34.37   24.39
      32768   48.17   49.14   48.23   37.27   47.21   36.27
      65536   58.34   58.49   58.34   44.17   58.91   46.39
     131072   66.17   65.41   66.57   51.46   66.18   54.81
     262144   70.92   69.49   70.72   67.99   70.54   67.33
     524288   73.05   71.48   72.73   71.05   73.21   70.33
    1048576   74.23   72.51   73.83   71.44   73.88   71.50
    2097152   74.58   73.11   74.51   71.72   74.41   71.66
    4194304   74.63   73.44   74.18   72.73   74.17   72.69
- The 'less good':
Although the disk claims to be a drop in replacement for old SCA HDDs, I had to spend quite a bit of time tinkering with the jumpers on the drive to reach a semi-stable configuration that would boot or reboot right 100% of the time.
Here's the configuration that I am using:
- Debug set to 0x02 on the Tezro (with 0x0 and upon cold boot I was often encountering SCSI timeouts and resets). With 0x02 it's a little better.
- Jumper set on JBD2 - pins 15-16 (a notch down from the fastest speed, according to Red Rock)

- The 'currently being investigated':

When booting/rebooting the Tezro with the drive alone on the ql0 bus, I still get random SCSI timeouts/resets but once booted, everything is fine and stable.
I am still working with Red Rock Tech Support to figure out what's going on. This is a fully up-to-date Tezro with 6.5.30f + patches. PROM (6.211) and L1 (1.48.1) up-to-date
This is what it looks like on the serial console:

- Example of a reboot with only one SCSI reset:
Code:
Selecting IO9 baseio
ql0: Command timeout, draining ctlr IO for as long as 10 seconds
ql0d1: SCSI command timeout: 1 commands: 0x0 
ql0: Resetting SCSI bus.
NOTICE: usb_hid_attach() reattach of device on bus 0xa800000001479740 path 2
CPU 2: hwgraph_edge_add() failed.  src vhdl=0xfb dest vhdl=0xfd name=mouse
NOTICE: Starting failsoftd
[...]
The system is coming up.

- Example of a reboot with multiple
Code:
Selecting IO9 baseio
ql0: Command timeout, draining ctlr IO for as long as 10 seconds
ql0d1: SCSI command timeout: 1 commands: 0x0 
ql0: Resetting SCSI bus.
ql0: Command timeout, draining ctlr IO for as long as 10 seconds
ql0d1: SCSI command timeout: 1 commands: 0x0 
ql0: Resetting SCSI bus.
ql0: Command timeout, draining ctlr IO for as long as 10 seconds
ql0d1: SCSI command timeout: 1 commands: 0x0 
ql0: Resetting SCSI bus.
NOTICE: usb_hid_attach() reattach of device on bus 0xa800000001479740 path 2
CPU 2: hwgraph_edge_add() failed.  src vhdl=0xfb dest vhdl=0xfd name=mouse
NOTICE: Starting failsoftd
[...]
The system is coming up.
On a few occasions (cold boot) I got xbrick messages and the system went back to the PROM menu. Then starting it from there (option #1) usually worked fine.

Here are a few pictures I took:
IMG_20221109_212816.jpg

IMG_20221109_214710.jpg


IMG_20221109_213323.jpg
 
  • Like
Reactions: Elf and flexion

mgtremaine

Member
May 3, 2020
62
35
18
San Diego, Ca
www.stellarcore.net
Thanks for sharing this. I think many of us have seen these drives and wondered. The price for the hobby market is obviously really high. Would love to see some sort of 18/32GB offering from them and something more affordable for us (especially the o2/Octane users) but probably just a dream.

Let's us know how this resolves.

-Mike
 

Elf

Storybook / Retired, ex-staff
Feb 4, 2019
792
252
63
Mountain West (US)
Thanks for posting! I remember seeing the challenges with this in the Discord chat; quite the investment for the Tezro, hopefully it stabilizes! :)
 

ElCoyote_

New member
May 25, 2022
8
9
3
I agree, it would be great if they could just sell the bare enclosure. So far, the list of successfully tested drives is this:
- Intel DS3-S4510
- Samsung EVO 840 500gb
- Samsung EVO 870 4Tb (tested with 6.5.30f and patches and everything was fine but of course my /dev/root needed to be < 1Tb)
I have a 1.82Tb DS3-S4610 from ebay that's incoming and which I will use to replace the S4510 (it's read-intensive while the S4610 is 'mixed' and better at writes).
 
  • Like
Reactions: Elf

ElCoyote_

New member
May 25, 2022
8
9
3
I agree, for the hobbyist it is way too much.. I had to think twice before buying just one drive but at the end of the day I have a single box which needs it and it's my beloved Tezro so I ended up getting a drive for it.
 

ElCoyote_

New member
May 25, 2022
8
9
3
I have some good news.
Yesterday Red Rock Engineering identified an improvement in their firmware and sent it to me.
I updated the Red Rock drive using USB + Debugging cable from a Windows 10 laptop and continued testing.

When I next booted the Tezro (cold from the night), I did not see a single SCSI timeout.
Over the next 4 warm reboots I only got a single SCSI timeout -once-.
When next I did a cold reboot, I got 3 SCSI timeouts but the whole situation seems better overall.


[root@istar ~]# diskpatch -v
sc0d1l0: Disk Red RockSTXS B00C <==== STANDARD FIRMWARE
sc1d2l0: Disk ZULUSCSIHARDDRIVE 1.1 Serial: B60090A2

[root@istar ~]# diskpatch -v
sc0d1l0: Disk Red RockSTXS B00x <==== ENGINEERING DEBUG FIRMWARE
sc1d2l0: Disk ZULUSCSIHARDDRIVE 1.1 Serial: B60090A2
 
Last edited:
  • Like
Reactions: Elf

ElCoyote_

New member
May 25, 2022
8
9
3
So this afternoon the drive came back from Red Rock. There were no issues during their internal testing (including temperature testing) and they sent it back, albeit with a newer firmware.

I've now been hammering the drive for 3 hours, tested two SSDs (Intel D3-S4510 and D3-4610), removed the JB2 15-16 I was previously using for stability and I have not seen a single SCSI reset.

so Kudos to the Red Rock team for working on this with me (I had purchased only a -SINGLE- drive from them). I'm very impressed and humbled.

Code:
# diskpatch -v
sc0d1l0:  Disk         Red RockSTXS            B01D  Serial: PHYG2085
sc1d2l0:  Disk         ZULUSCSIHARDDRIVE       1.1   Serial: B60090A2
Which gave me (with my D3-S4610 1.82Tb drive in the enclosure):

Code:
[root@istar ~]#  diskperf -D -W -n "Red Rock SCA + Intel DS3 S4610 1.82Tb" -r4k /d/test
#---------------------------------------------------------
# Disk Performance Test Results Generated By Diskperf V1.2
#
# Test name     : Red Rock SCA + Intel DS3 S4610 1.82Tb
# Test date     : Tue Dec 13 17:59:37 2022
# Test machine  : IRIX64 istar 6.5 07202013 IP35
# Test type     : XFS data subvolume
# Test path     : /d/test
# Request sizes : min=4096 max=4194304
# Parameters    : direct=1 time=10 scale=1.000 delay=0.000
# XFS file size : 466583552 bytes
#---------------------------------------------------------
# req_size  fwd_wt  fwd_rd  bwd_wt  bwd_rd  rnd_wt  rnd_rd
#  (bytes)  (MB/s)  (MB/s)  (MB/s)  (MB/s)  (MB/s)  (MB/s)
#---------------------------------------------------------
       4096   13.20   14.49   13.48   13.28   13.28   10.47
       8192   23.19   24.26   22.37   21.96   22.93   17.43
      16384   35.85   36.99   35.71   24.63   35.66   26.28
      32768   48.39   48.81   48.36   43.84   47.98   37.42
      65536   57.82   58.70   59.18   53.65   58.52   49.03
     131072   66.05   65.77   66.55   64.29   66.24   58.10
     262144   70.74   69.66   70.56   67.85   70.57   67.41
     524288   72.58   71.73   72.60   71.25   73.24   70.41
    1048576   74.30   72.68   74.04   71.67   73.61   71.74
    2097152   74.41   73.20   74.05   71.72   74.06   71.93
    4194304   74.28   73.49   74.34   72.86   74.88   72.79
 
  • Like
Reactions: Elf

About us

  • Silicon Graphics User Group (SGUG) is a community for users, developers, and admirers of Silicon Graphics (SGI) products. We aim to be a friendly hobbyist community for discussing all aspects of SGIs, including use, software development, the IRIX Operating System, and troubleshooting, as well as facilitating hardware exchange.

User Menu