HP Proliant Microserver broken SB700/SB800 driver?

W

willyvmm

HI.

I have a problem with SB Driver (I think).

The system is Nas4Free.
I have tried many different kernel options/drivers, at least I Used GENERIC Kernel from FreeBSD-9.1-RC1-amd64-memstick.img installation image.

There is a System hdd and 4 data disks connected to SAS controller.
Disks connected to SAS are working fine, but there is a problem with system disk.
When I try to do something like umount/mount/newfs etc, even copy ... The system hang up for some seconds with one of the following messages:

Code: 
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 ab 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 ac 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 7f 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 ac 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 7f 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 7f 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 ac 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
(ada4:ata0:0:1:0): READ_DMA. ACB: c8 00 9a f5 7f 40 00 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 7e 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command

I have tried to remove everything from loader.conf, rc.conf etc... no luck.
For example newfs is working fine 2-3 times, and after that errors are coming.

HDD is 100% ok. I have replaced 3 different Hard disks from different manufacturers. S.M.A.R.T is clean and in good condition.

Any idea how to fix it ?
 
I'm using exactly the same hardware. But, the CD is connected to the eSATA connector, and the system disk (2,5" Fujitsu) is connected to the onboard SATA and 4x 2T Seagate connected to SAS.
 
You can try flashing a patched Microserver BIOS. It enables AHCI on ports 5,6. I flash it because SSD on ODD port runs with errors and timeouts. Google "microserver patched BIOS".
 
Thanks @borov that solved the problem. I've made some quick tests and it seems that the problem is gone.

Virtual BEER for you :beergrin

Anyway, I still think the driver is broken because is not working correctly with the original BIOS with the strange "SATA combined mode".

Best Regards
Willy.
 
Last edited by a moderator:
Same Error Messages

willyvmm said:
When I try to do something like umount/mount/newfs etc, even copy ... The system hang up for some seconds with one of the following messages:

Code: 
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 ab 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
...
(ada4:ata0:0:1:0): READ_DMA48. ACB: 25 00 7e 9f a1 40 11 00 00 00 01 00
(ada4:ata0:0:1:0): CAM status: Command timeout
(ada4:ata0:0:1:0): Retrying command
Click to expand...

Exactly the same error messages occurred on my HP Proliant Microserver NL40 while the system was scrubbing the ZFS pool or during an RSYNC process. The harddisk is connected via the eSATA port.

borov said:
You can try flash patched Microserver BIOS.
Click to expand...

Thanks for the solution! I've finally got rid of these errors.
 
willyvmm said:
There is a System hdd and 4 data disks connected to SAS controller.
Click to expand...

FYI, it's actually a SATA controller. It's just presented through what's usually a SAS connector (SFF-8087) probably to save space on the small motherboard.


The mind boggles as to why HP decided to use combined mode on the MicroServer southbridge. Given that one of the two affected ports is an eSATA one where hotplugging would be a normal requirement, crippling the port into legacy IDE compatibility mode just makes no sense at all.

There are two modified BIOS's out there that I've seen, one by "Diehard" and one by "TheBay":

Diehard's one exposes many, many hidden chipset settings most of which are unrelated to SATA and I feel it's probably risky to use in case you change something that impacts the stability of the machine.

TheBay's one only enables switching the combined mode off so is simpler and probably less risky to use.

I started with the first one but quickly switched over to TheBay's one and have had no issues so far.
 
I have a pair of G7s on the way. Does anyone know if the latest HP BIOS fixes this problem or do you have to still use a patched BIOS to get full AHCI functionality?
 
You must log in or register to reply here.
Back
Top