Re: poor raidframe performance while setting it up

[Bernd Ernesti <netbsd%lists.veego.de@localhost>]

On Sun, Jun 07, 2009 at 03:57:39PM -0400, Matthew Mondor wrote:
> On Sun, 7 Jun 2009 21:12:18 +0200
> Bernd Ernesti <netbsd%lists.veego.de@localhost> wrote:
> 
> > First it seems to be fast with 75MB/s i/o on both drives and then it
> > slows down very dramatically to 2.5MB/s, so the initail would take 100
> > till 200 hours for this 1TB raid set. With 75MB/s it would have taken
> > 3-4 hours.
> > The slow down happens after a few seonds till 2-3 minutes.
> 
> Sorry if this is too obvious, but it's worth making sure of; smart
> statistics for both drives show no reallocated sectors?  I've seen
> performance drop a lot in those cases, even though the drives continued
> to work temporarily (but need fast replacement), and I'd assume raid
> would have to slow down at the speed of the slowest drive in most
> cases.  Of course, performance would also drop somewhere when using
> the disk without raidframe.

I checked that first and the counter is still zero.
A short smart self test doesn't produce any errors either.
And I also tried to dd to and from both drives at the same time after I
encountered the performance problem. There were no problem.

I also swapped both drives, so the faster one is now wd2, but the
same problem occurs with wd3.

> Also, no interesting messages in dmesg like autocorrected software
> errors or the like?

No :(

> The name of the controller might also help if a known problem exists
> with a particular device or driver...  Some controllers also have
> performance problems when operating on more than one device (although
> this performance drop appears dramatic).

ahcisata0 at pci0 dev 17 function 0: ATI Technologies SB700/SB800 SATA 
Controller (AHCI mode)
ioapic0: int22 0xa064<vector=0x64,delmode=0x0,actlo,level,dest=0x0> 
0x0<target=0x0>
ahcisata0: interrupting at ioapic0 pin 22
ahcisata0: AHCI revision 1.1, 6 ports, 32 command slots, features 0xf722e080
ahcisata0 port 2: device present, speed: 3.0Gb/s
ahcisata0 port 3: device present, speed: 3.0Gb/s
wd2 at atabus2 drive 0: <WDC WD1000FYPS0>
wd2: drive supports 16-sector PIO transfers, LBA48 addressing
wd2: 931 GB, 1938021 cyl, 16 head, 63 sec, 512 bytes/sect x 1953525168 sectors
wd2: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133)
wd2(ahcisata0:2:0): using PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133) 
(using DMA)
wd3 at atabus3 drive 0: <WDC WD10EADS>
wd3: drive supports 16-sector PIO transfers, LBA48 addressing
wd3: 931 GB, 1938021 cyl, 16 head, 63 sec, 512 bytes/sect x 1953525168 sectors
wd3: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133)
wd3(ahcisata0:3:0): using PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133) 
(using DMA)

Bernd