1 Rookie
•
2 Posts
0
46
February 22nd, 2025 02:31
RAID Controller Misidentifying SSDs as Rotational on Server
If there's a RAID expert here, we would appreciate your assistance with the following.
We are experiencing an issue with our server Dell PowerEdge R630 running RAID 10 on PERC H730 Mini RAID controller with Ubuntu 22 Lts.
Our recent performance tests using FIFO commands show read speeds of 2700 MB/s and write speeds of 1800 MB/s, which are lower than expected for our RAID-10 setup.
Upon investigation, we found that the Dell PERC H730 Mini RAID controller (Broadcom MegaRAID SAS-3 3108) is incorrectly detecting our SATA SSDs (Micron 5210) as rotational devices (ROTA=1) instead of non-rotational (ROTA=0). This misclassification can lead to suboptimal I/O scheduling, affecting overall performance.
The output of lsblk -o NAME,ROTA,DISC-MAX
So on recommendations we added a file
/etc/udev/rules.d/99-non-rotational.rules
Added the content
ACTION=="add|change", ENV{DM_NAME}=="ubuntu--vg-ubuntu--lv", ATTR{queue/rotational}="0"
ACTION=="add|change", ENV{DM_NAME}=="ubuntu--vg-lv--kvmdata", ATTR{queue/rotational}="0"
ACTION=="add|change", ENV{DM_NAME}=="ubuntu--vg-vsv1017--dn4S8Q51DQNaT7fg--3bRpfjxjIRtCg5r3", ATTR{queue/rotational}="0"
ACTION=="add|change", ENV{DM_NAME}=="ubuntu--vg-vsv1018--dejnhTUJQifyOUDT--tKQdhCLEotCGP7mz", ATTR{queue/rotational}="0"
ACTION=="add|change", ENV{DM_NAME}=="ubuntu--vg-vsv1019--dPSXM0StizWx1RB5--vZ58gXA2ryO3ksqQ", ATTR{queue/rotational}="0"
ACTION=="add|change", ENV{DM_NAME}=="ubuntu--vg-vsv1027--d4HiXQkMj1CoCoIZ--AU8xHN14C7U9vYkA", ATTR{queue/rotational}="0"
ACTION=="add|change", KERNEL=="sd[a-z]", ATTR{queue/rotational}="0"
executed the command
udevadm control --reload-rules
udevadm trigger
update-initramfs -u # So that the setting remain intact on reboot
The output of lsblk -o NAME,ROTA,DISC-MAX after the changes
And the speed test results
fio --name=mixed_rw50 --ioengine=libaio --rw=randrw --rwmixread=50 --bs=4k --numjobs=4 --size=1G --runtime=60 --group_reporting --filename=/tmp/fio_test
mixed_rw50: (g=0): rw=randrw, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=1
...
fio-3.28
Starting 4 processes
Jobs: 4 (f=4): [m(4)][64.3%][r=402MiB/s,w=403MiB/s][r=103k,w=103k IOPS][eta 00m:05s]
mixed_rw50: (groupid=0, jobs=4): err= 0: pid=3142982: Sat Feb 22 07:55:55 2025
read: IOPS=54.2k, BW=212MiB/s (222MB/s)(2046MiB/9659msec)
slat (nsec): min=1546, max=8143.4k, avg=59369.06, stdev=121660.94
clat (nsec): min=747, max=1128.5k, avg=1361.98, stdev=2555.22
lat (usec): min=2, max=8150, avg=60.91, stdev=122.25
clat percentiles (nsec):
| 1.00th=[ 764], 5.00th=[ 932], 10.00th=[ 932], 20.00th=[ 940],
| 30.00th=[ 948], 40.00th=[ 956], 50.00th=[ 980], 60.00th=[ 1256],
| 70.00th=[ 1400], 80.00th=[ 1576], 90.00th=[ 2256], 95.00th=[ 2800],
| 99.00th=[ 4048], 99.50th=[ 5088], 99.90th=[11072], 99.95th=[13632],
| 99.99th=[19584]
bw ( KiB/s): min=25328, max=962488, per=93.04%, avg=201791.53, stdev=55412.08, samples=76
iops : min= 6332, max=240622, avg=50447.68, stdev=13853.02, samples=76
write: IOPS=54.3k, BW=212MiB/s (223MB/s)(2050MiB/9659msec); 0 zone resets
slat (nsec): min=1949, max=2050.3k, avg=8387.45, stdev=39394.09
clat (nsec): min=751, max=1162.7k, avg=1130.97, stdev=1715.29
lat (usec): min=2, max=2054, avg= 9.64, stdev=39.52
clat percentiles (nsec):
| 1.00th=[ 764], 5.00th=[ 812], 10.00th=[ 940], 20.00th=[ 940],
| 30.00th=[ 948], 40.00th=[ 956], 50.00th=[ 964], 60.00th=[ 988],
| 70.00th=[ 1064], 80.00th=[ 1336], 90.00th=[ 1464], 95.00th=[ 1656],
| 99.00th=[ 2768], 99.50th=[ 3312], 99.90th=[ 9408], 99.95th=[12096],
| 99.99th=[17792]
bw ( KiB/s): min=26480, max=966720, per=92.97%, avg=202056.89, stdev=55586.49, samples=76
iops : min= 6620, max=241680, avg=50514.05, stdev=13896.64, samples=76
lat (nsec) : 750=0.01%, 1000=57.62%
lat (usec) : 2=34.98%, 4=6.70%, 10=0.59%, 20=0.10%, 50=0.01%
lat (usec) : 100=0.01%, 250=0.01%, 1000=0.01%
lat (msec) : 2=0.01%
cpu : usr=7.05%, sys=24.70%, ctx=131511, majf=0, minf=177
IO depths : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0%
submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0%
issued rwts: total=523742,524834,0,0 short=0,0,0,0 dropped=0,0,0,0
latency : target=0, window=0, percentile=100.00%, depth=1
Run status group 0 (all jobs):
READ: bw=212MiB/s (222MB/s), 212MiB/s-212MiB/s (222MB/s-222MB/s), io=2046MiB (2145MB), run=9659-9659msec
WRITE: bw=212MiB/s (223MB/s), 212MiB/s-212MiB/s (223MB/s-223MB/s), io=2050MiB (2150MB), run=9659-9659msec
We would appreciate help from anyone who has encountered this issue and found a workaround to improve performance, which we believe can reach up to 6GBps to 7GBps.
Thank you in advance!
DELL-Young E
Moderator
•
5.1K Posts
0
February 24th, 2025 01:38
Hello, thanks for choosing Dell.
I believe this is more for the consulting realm than troubleshooting that us tech support offer.
What I see is that SATA SSDs (Micron 5210) -> I don’t think this is a Dell part
This read may help:
https://dell.to/4gWYDwm
You could try tweaking Read Policy to No-Read-Ahead.
Respectfully,