Re: Perc 5/i SAS

[Havard Eidnes <he%uninett.no@localhost>]

> On Thu, Aug 16, 2012 at 09:45:49AM +0200, moda wrote:
>> Netbsd 5.1.2 x64 = 13MB/s
> 
> That is abysmal, something must be wrong in your setup - can you provide
> more details (dmesg, raid configuration, file system mount options [the
> line from /etc/fstab for your target filesystem])?

My gut reaction is that there must be some missing parallelism in
the NetBSD case, e.g. that the number of "openings" (which limits
the parallelism) is set way too low.  That's happened before with
other raid controller drivers if I recall correctly.  I could not
find any "smoking gun" in the CVS history log for mfi.c, though.

If I test a new machine locally running 6.0_BETA2 doing a large
sequential file write, I get:

dd if=/dev/zero of=bigfile bs=1m count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 11.907 secs (88063828 bytes/sec)

and

      tty              sd0               sd1               sd2             CPU
 tin tout  KB/t  t/s  MB/s   KB/t  t/s  MB/s   KB/t  t/s  MB/s  us ni sy in id
   3   83 6.000    2 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   0  239 0.000    0 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   1   81 0.000    0 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   4  138 7.778    9 0.068  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   0   80 0.000    0 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   2   84 0.000    0 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   1   92 0.000    0 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   4   84 0.000    0 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   6  132 61.94   31 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   0   80 63.52 1937 120.1  0.000    0 0.000  0.000    0 0.000   0  0  4  0 96
   0   79 60.11  373 21.91  0.000    0 0.000  0.000    0 0.000   0  0  1  0 99
   0   79 63.41 2130 131.9  0.000    0 0.000  0.000    0 0.000   0  0  4  0 95
   0   79 61.90  182 11.01  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   0   80 63.11 1023 63.03  0.000    0 0.000  0.000    0 0.000   0  0  2  0 98
   0   79 63.80 2210 137.7  0.000    0 0.000  0.000    0 0.000   0  0  5  0 95
   0   79 63.84 1346 83.88  0.000    0 0.000  0.000    0 0.000   0  0  2  0 98
   0   79 63.76 1388 86.43  0.000    0 0.000  0.000    0 0.000   0  0  3  0 97
   0   79 63.77 1421 88.47  0.000    0 0.000  0.000    0 0.000   0  0  3  0 97
   0   79 63.81 1175 73.24  0.000    0 0.000  0.000    0 0.000   0  0  2  0 98
   0   79 63.79 1614 100.5  0.000    0 0.000  0.000    0 0.000   0  0  3  0 96
      tty              sd0               sd1               sd2             CPU
 tin tout  KB/t  t/s  MB/s   KB/t  t/s  MB/s   KB/t  t/s  MB/s  us ni sy in id
   0  195 63.40 1150 71.17  0.000    0 0.000  0.000    0 0.000   0  0  2  0 98
   0  238 14.65  128 1.827  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   0   80 0.000    0 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100
   0   80 0.000    0 0.000  0.000    0 0.000  0.000    0 0.000   0  0  0  0 100

This is with a

mfi0 at pci1 dev 0 function 0: SAS 9260-8i
mfi0: interrupting at ioapic0 pin 18
mfi0: logical drives 3, version 12.12.0-0111, 512MB RAM
scsibus0 at mfi0: 64 targets, 8 luns per target

Regards,

- Håvard