Extremely poor disk performance on ProLiant DL360 G5 RAID

[Frank Wille <frank%phoenix.owl.de@localhost>]

Hi,

I'm currently setting up a HP ProLiant DL360 G5 with NetBSD/amd64 8.0RC2.
It has 16GB RAM...
NetBSD 8.0_RC2 (GENERIC) #0: Tue Jul  3 07:13:41 UTC 2018
        mkrepro%mkrepro.NetBSD.org@localhost:/usr/src/sys/arch/amd64/compile/GENERIC
total memory = 16381 MB
avail memory = 15883 MB

...Xeon 5160 3GHz (2 cores with hyperthreading, I guess)
cpu0 at mainbus0 apid 0
cpu0: Intel(R) Xeon(R) CPU            5160  @ 3.00GHz, id 0x6f6
cpu0: package 0, core 0, smt 0
cpu1 at mainbus0 apid 6
cpu1: Intel(R) Xeon(R) CPU            5160  @ 3.00GHz, id 0x6f6
cpu1: package 3, core 0, smt 0
cpu2 at mainbus0 apid 1
cpu2: Intel(R) Xeon(R) CPU            5160  @ 3.00GHz, id 0x6f6
cpu2: package 0, core 1, smt 0
cpu3 at mainbus0 apid 7
cpu3: Intel(R) Xeon(R) CPU            5160  @ 3.00GHz, id 0x6f6
cpu3: package 3, core 1, smt 0

...and two 550 GB SAS RAID-1 disks, configured via the BIOS and appearing
as a single SCSI disk:
ciss0 at pci7 dev 0 function 0: HP Smart Array 3
ciss0: interrupting at ioapic0 pin 16
ciss0: 1 LD, HW rev 3, FW 5.20/5.20, 64bit fifo
scsibus0 at ciss0: 1 target, 1 lun per target
[...]
sd0 at scsibus0 target 0 lun 0: <HP, LOGICAL VOLUME, 5.20> disk fixed
sd0: 558 GB, 65535 cyl, 255 head, 70 sec, 512 bytes/sect x 1172058032 sectors
sd0: async, 8-bit transfers, tagged queueing

I'm wondering about the extremely slow disk performance while extracting
several pkgsrc archives. Especially the boost-jam (40 minutes) and
boost-headers (35 minutes) archives take annoyingly long.

The system is still in a freshly installed state with only default services
running. The load is constantly 1.00 while extracting, but I don't see
where the processing power is lost. bzcat and tar only use around 8%.

load averages:  1.00,  1.00,  0.92;               up 5+21:35:12        11:37:06
32 processes: 30 sleeping, 2 on CPU
CPU states:  2.4% user,  0.0% nice,  0.2% system,  0.0% interrupt, 97.2% idle
Memory: 4622M Act, 212M Inact, 276K Wired, 14M Exec, 4795M File, 9833M Free
Swap: 16G Total, 16G Free

  PID USERNAME PRI NICE   SIZE   RES STATE      TIME   WCPU    CPU COMMAND
16121 root      85    0    14M 4400K pipe_w/2   0:05  5.32%  5.32% bzcat
 9938 root      85    0    10M 1220K biowai/0   0:06  2.83%  2.83% tar
    0 root       0    0     0K  102M CPU/3     24:25  0.00%  0.00% [system]
 6716 frank     85    0    82M 4888K select/1   0:06  0.00%  0.00% sshd
  571 root      85    0    48M 2420K kqueue/3   0:01  0.00%  0.00% master
[...]

No interrupt storm either:

interrupt          total rate
TLB shootdown    5945470   11
cpu0 timer      51075258   99
ioapic0 pin 1       1643    0
ioapic0 pin 16   3277090    6
ioapic0 pin 18   6109904   11
ioapic0 pin 22        48    0
ioapic0 pin 14        12    0
Total           66409425  129

Anybody knows why the disk operations on this servers are so slow?
What can I check?

-- 
Frank Wille