Post Your HDD performance info

[antik]

# diskinfo -t /dev/ad0
/dev/ad0
512 # sectorsize
160041885696 # mediasize in bytes (149G)
312581808 # mediasize in sectors
310101 # Cylinders according to firmware.
16 # Heads according to firmware.
63 # Sectors according to firmware.

Seek times:
Full stroke: 250 iter in 5.666521 sec = 22.666 msec
Half stroke: 250 iter in 4.371893 sec = 17.488 msec
Quarter stroke: 500 iter in 6.955999 sec = 13.912 msec
Short forward: 400 iter in 2.223611 sec = 5.559 msec
Short backward: 400 iter in 1.178749 sec = 2.947 msec
Seq outer: 2048 iter in 0.209462 sec = 0.102 msec
Seq inner: 2048 iter in 0.187843 sec = 0.092 msec
Transfer rates:
outside: 102400 kbytes in 1.844597 sec = 55513 kbytes/sec
middle: 102400 kbytes in 2.167990 sec = 47233 kbytes/sec
inside: 102400 kbytes in 3.464977 sec = 29553 kbytes/sec

# dmesg | grep ad0
ad0: 152627MB <Seagate ST3160023A 8.01> at ata0-master UDMA100

[madman]

/dev/ad0
512 # sectorsize
164696555520 # mediasize in bytes (153G)
321672960 # mediasize in sectors
5105920 # Cylinders according to firmware.
1 # Heads according to firmware.
63 # Sectors according to firmware.

Seek times:
Full stroke: 250 iter in 4.928724 sec = 19.715 msec
Half stroke: 250 iter in 3.628733 sec = 14.515 msec
Quarter stroke: 500 iter in 5.959027 sec = 11.918 msec
Short forward: 400 iter in 2.799953 sec = 7.000 msec
Short backward: 400 iter in 1.797054 sec = 4.493 msec
Seq outer: 2048 iter in 0.274963 sec = 0.134 msec
Seq inner: 2048 iter in 0.270974 sec = 0.132 msec
Transfer rates:
outside: 102400 kbytes in 1.728793 sec = 59232 kbytes/sec
middle: 102400 kbytes in 2.076237 sec = 49320 kbytes/sec
inside: 102400 kbytes in 3.553362 sec = 28818 kbytes/sec

[DrJ]

OK, I'll play...

diskinfo -t /dev/da0
512 # sectorsize
36703934464 # mediasize in bytes (34G)
71687372 # mediasize in sectors
4462 # Cylinders according to firmware.
255 # Heads according to firmware.
63 # Sectors according to firmware.

Seek times:
Full stroke: 250 iter in 2.168836 sec = 8.675 msec
Half stroke: 250 iter in 1.735579 sec = 6.942 msec
Quarter stroke: 500 iter in 2.827450 sec = 5.655 msec
Short forward: 400 iter in 1.745069 sec = 4.363 msec
Short backward: 400 iter in 0.826289 sec = 2.066 msec
Seq outer: 2048 iter in 0.372151 sec = 0.182 msec
Seq inner: 2048 iter in 0.363974 sec = 0.178 msec
Transfer rates:
outside: 102400 kbytes in 1.383251 sec = 74029 kbytes/sec
middle: 102400 kbytes in 1.556284 sec = 65798 kbytes/sec
inside: 102400 kbytes in 2.113208 sec = 48457 kbytes/sec

pinot# dmesg | grep da0
da0 at sym0 bus 0 target 1 lun 0
da0: <SEAGATE ST336753LW 0005> Fixed Direct Access SCSI-3 device
da0: 160.000MB/s transfers (80.000MHz, offset 62, 16bit), Tagged Queueing Enabled

diskinfo -t /dev/da1
/dev/da1
512 # sectorsize
146815737856 # mediasize in bytes (137G)
286749488 # mediasize in sectors
17849 # Cylinders according to firmware.
255 # Heads according to firmware.
63 # Sectors according to firmware.

Seek times:
Full stroke: 250 iter in 3.100419 sec = 12.402 msec
Half stroke: 250 iter in 2.289167 sec = 9.157 msec
Quarter stroke: 500 iter in 3.875068 sec = 7.750 msec
Short forward: 400 iter in 1.309084 sec = 3.273 msec
Short backward: 400 iter in 1.287727 sec = 3.219 msec
Seq outer: 2048 iter in 0.375318 sec = 0.183 msec
Seq inner: 2048 iter in 0.371993 sec = 0.182 msec
Transfer rates:
outside: 102400 kbytes in 1.513601 sec = 67653 kbytes/sec
middle: 102400 kbytes in 1.838785 sec = 55689 kbytes/sec
inside: 102400 kbytes in 2.727111 sec = 37549 kbytes/sec

pinot# dmesg | grep da1
da1 at sym0 bus 0 target 2 lun 0
da1: <SEAGATE ST3146807LW 0006> Fixed Direct Access SCSI-3 device
da1: 160.000MB/s transfers (80.000MHz, offset 62, 16bit), Tagged Queueing Enabled
da1: 140014MB (286749488 512 byte sectors: 255H 63S/T 17849C)


DrJ

[Puffball]

%diskinfo -t /dev/ad0
/dev/ad0
512 # sectorsize
80026361856 # mediasize in bytes (75G)
156301488 # mediasize in sectors
155061 # Cylinders according to firmware.
16 # Heads according to firmware.
63 # Sectors according to firmware.

Seek times:
Full stroke: 250 iter in 5.696677 sec = 22.787 msec
Half stroke: 250 iter in 4.150914 sec = 16.604 msec
Quarter stroke: 500 iter in 7.032293 sec = 14.065 msec
Short forward: 400 iter in 3.514182 sec = 8.785 msec
Short backward: 400 iter in 2.055613 sec = 5.139 msec
Seq outer: 2048 iter in 0.314303 sec = 0.153 msec
Seq inner: 2048 iter in 0.304994 sec = 0.149 msec
Transfer rates:
outside: 102400 kbytes in 2.134657 sec = 47970 kbytes/sec
middle: 102400 kbytes in 2.279191 sec = 44928 kbytes/sec
inside: 102400 kbytes in 3.168058 sec = 32323 kbytes/sec

I think the transfer rate is lower than it should be, I'm sure its not my cable's fault why its in UDMA 33 mode, but I think the lack of a proprietary driver is preventing it from using UDMA 66/100; not sure if any BSD tweak could help this.

[DrJ]

It would help if you would post the output from dmesg about your drive. At first blush it looks OK to me if it is a few year old IDE drive. It is already transferring data faster than UDMA33, so I don't think that's the issue.

DrJ

[antik]

I think the transfer rate is lower than it should be, I'm sure its not my cable's fault why its in UDMA 33 mode, but I think the lack of a proprietary driver is preventing it from using UDMA 66/100; not sure if any BSD tweak could help this.
Today I found that my CoolerMaster 133Mbit rounded cable is faulty and I forced to change it to 80pin flat one.
Before:
ad0: DMA limited to UDMA33, device found non-ATA66 cable
ad0: 152627MB <Seagate ST3160023A 8.01> at ata0-master UDMA33

After:
ad0: 152627MB <Seagate ST3160023A 8.01> at ata0-master UDMA100

[Puffball]

I guess I was primarily looking at transfer rate.

%dmesg | grep ad0
ad0: 76319MB <WDC WD800JB-00FMA0 13.03G13> at ata0-master UDMA100
Trying to mount root from ufs:/dev/ad0s1a

Yeah, its fine. @_@

[DrJ]

They are a tad slow for modern drives, but they should be fine. Don't compare your transfer rates with the ones I posted. Both are SCSI drives; the first one is a 15K RPM and the second one is a 10K. Those are different beasts than a 7.2K RPM ATA drive.

It would also help if you would edit your post and use the "code" delimiter. Using "quote" messes up the formatting (I found this out the hard way).

DrJ

[Puffball]

Ah, does 'code' come installed by default with phpBB, you think? It sounds kind of niche/tech-centric.

[DrJ]

"Code" is the button to the right of "quote" in the posting panel. Select your text, and then hit the code button. Give it a try, and have a look at the preview panel (button on the bottom).

DrJ

[Puffball]

Oh, I know what it is, it just strikes me as technical-specific, what's a 'Lesbian and Toy Molding' community supposed to do with a "Code" button? ;P

[DrJ]

:)

[TerryP]

I'll post the one from my personal machine if I can ever get PCBSD installed again.

[DrJ]

I thought I would add one more drive test that shows bus saturation. This one is also a 15K RPM SCSI drive, but it is on a very old computer that has a slow disk controller (a Sparc-64 on FreeBSD 6.0). The seek times are the same as the one I posted above; the transfer rates, however, are constant across the platter. That's a good indication that the bus limits the speed of the drive.

For my purposes of serving static web pages, it is good enough, as it can fully load its single 100BaseT Ethernet port. It can't do a lot more, though.

shiraz# !96
diskinfo -t /dev/da0
/dev/da0
512 # sectorsize
18351967232 # mediasize in bytes (17G)
35843686 # mediasize in sectors
2231 # Cylinders according to firmware.
255 # Heads according to firmware.
63 # Sectors according to firmware.

Seek times:
Full stroke: 250 iter in 2.217195 sec = 8.869 msec
Half stroke: 250 iter in 1.764256 sec = 7.057 msec
Quarter stroke: 500 iter in 2.978984 sec = 5.958 msec
Short forward: 400 iter in 1.684741 sec = 4.212 msec
Short backward: 400 iter in 1.058240 sec = 2.646 msec
Seq outer: 2048 iter in 0.630573 sec = 0.308 msec
Seq inner: 2048 iter in 0.624827 sec = 0.305 msec
Transfer rates:
outside: 102400 kbytes in 6.807583 sec = 15042 kbytes/sec
middle: 102400 kbytes in 6.825817 sec = 15002 kbytes/sec
inside: 102400 kbytes in 6.857502 sec = 14933 kbytes/sec

shiraz# dmesg | grep da0
da0 at esp0 bus 0 target 0 lun 0
da0: <SEAGATE ST318453LC 0005> Fixed Direct Access SCSI-3 device
da0: 20.000MB/s transfers (10.000MHz, offset 15, 16bit), Tagged Queueing Enabledda0: 17501MB (35843686 512 byte sectors: 255H 63S/T 2231C)


So why use a drive that is too fast for the controller? For this application, the computer has SCSI only (and no IDE card is available) and the drive is reliable and quiet. Plus, it was cheap.

DrJ