FreeBSD 10.2 + Adaptec RAID 5805?

Question

FreeBSD

Hello.
There is a server with fresh FreeBSD 10.2 and an Adaptec RAID 5805 controller with 4 SATA drives in JBOD.
dmesg

aac0: mem 0xf7e00000-0xf7ffffff irq 16 at device 0.0 on pci130
aac0: Enabling 64-bit address support
aac0: Enable Raw I/O
aac0: Enable 64-bit array
aac0: New comm. interface enabled
aac0: Adaptec 5805, aac driver 2.1.9-1
aacp0 on aac0
aacp1 on aac0
aacp2 on aac0

cam control devlist

< ST31000528AS CC37> at scbus8 target 0 lun 0 (pass1)
< ST31000528AS CC37> at scbus8 target 1 lun 0 (pass2)
< ST31000528AS CC37> at scbus8 target 2 lun 0 (pass3)
< ST31000528AS CC37> at scbus8 target 3 lun 0 ( pass4)

It seems that the controller is detected normally, but fdisk does not see the disks themselves and they are not in / dev

# ls /dev
aac0 consolectl cuau2.lock geom.ctl kmem pass0 sndstat ttyu1.lock ttyv6 ufssuspend ugen7.2
acpi ctty da0 gptid led pass1 stderr ttyu2 ttyv7 ugen0.1 ukbd0
afa0 cuau0 da0p1 hpet0 log pass2 stdin ttyu2.init 1ttyv1.18 ums0
apm cuau0.init da0p2 hpn0 mdctl pass3 stdout ttyu2.lock ttyv9 ugen2.1 urandom
apmctl cuau0.lock da0p3 io mem pass4 sysmouse ttyv0 ttyva ugen2.2 usb
atkbd0 cuau1 devctl kbd0 midistat pass5 ttyu0 ttyv1 ttyvb ugen3.1 usbctl
audit cuau1.init devstat kbd1 mixer0 pci ttyu0.init ttyv2 tty.1 dumpplock ugen14.1
dumpplock ugen kbd2 mixer1 pts ttyu0.lock ttyv3 ttyvd ugen5.1 zero
bpf0 cuau2 fd kbdmux0 nfslock random ttyu1 ttyv4 ttyve ugen6.1
console cuau2.init fido klog null ses0 ttyu1.init ttyv5 ttyvf ugen7.1

Tell me how to properly connect the disks? I work with BSD for the first time. In RHEL, everything started right away

Answer 1

Adaptec is evil, get rid of it immediately :-)
In JBOD mode, frya will not see disks, you need to create arrays on the controller. Then they will be visible. You can create arrays from one disk, and then build arrays of any configuration in software on a fra, on ZFS or on GEOM.
Adaptecs have one more unpleasant property - the firmware crashes there when a disk leaves the array. And the controller stops responding to requests altogether. We stumbled upon this and promptly replaced everything with LSI.