Drive Bay/Slot not matching on HP DL180 G6 H220

[ninnghazad]

Hey there!
I am trying to set up a used HP DL180 G6 using a HP H220 SAS HBA flashed to P20 LSI/BC-Firmware. While it seems to work, as in it installed and i can use the disks, i am still facing a rather weird issue:
The slots in the system are ordered like this:
0 2 4 6
1 3 5 7
And when i use sas2ircu to flash uid it works and the numbers used by sas2ircu fit the above mapping. However the disks don't. It's all crossed - a disk in physical slot 3 is seen by sas2ircu as being in slot 0. So whenever disk 2 gets used, the light for disk 1 flashes. I switched the SAS cables, but that just switches order on the 2 2x2 boxes - the weird crossing between actual slots/disks and leds/sas2ircu stays.
I remember the old controller (P410) that i ditched in favor of the H220 did not have this weird issue.
There is nothing on the disks yet, so i still can change stuff without risking data.

Don't know what is wrong, i don't even know if this is more of a cosmetic issue, or rather a symptom of something seriously wrong.
Maybe i flashed something wrong. The H220's bios nags me about it not being able to read some NVDATA, no idea what that's about.

sas2flash -list:

Code:

LSI Corporation SAS2 Flash Utility 
Version 16.00.00.00 (2013.03.01)  
Copyright (c) 2008-2013 LSI Corporation. All rights reserved  

		Adapter Selected is a LSI SAS: SAS2308_2(D1)  

		Controller Number			  : 0 
		Controller					 : SAS2308_2(D1)  
		PCI Address					: 00:06:00:00 
		SAS Address					: 500605b-0-09ce-fd40 
		NVDATA Version (Default)	   : 14.01.00.06 
		NVDATA Version (Persistent)	: 14.01.00.06 
		Firmware Product ID			: 0x2214 (IT) 
		Firmware Version			   : 20.00.07.00 
		NVDATA Vendor				  : LSI 
		NVDATA Product ID			  : SAS9207-8e 
		BIOS Version				   : 07.20.14.00 
		UEFI BSD Version			   : 07.27.01.01 
		FCODE Version				  : N/A 
		Board Name					 : H220 
		Board Assembly				 : H3-25278-05D 
		Board Tracer Number			: SV44738176

 

[ninnghazad]

so the bios having problems reading from nvdata was because it was the wrong bios version. i only flashed the uefi-bios, even though the dl180 g6 doesn´t do uefi...
after upgrading the bios it worked again, but now it seems the mainboard is somehow forcing all fans into "release-the-beast"-mode. great. so i erased the bios because i won´t boot from the card anyway (i had to use a freedos-disk with sas2flash to do so - bsd-sas2flash won´t do erase for me). i have freenas 11.1 on a usb-disk.
which means now i still have the crossed drives mapping, and additionally the system is kinda loud. annoying as it is meant for home use.
i read on the webs about lots of people having noise problems with the system, but for some flashing the hba to p20 seems to have resolved that. maybe the high IOH_TEMP (70deg C) is the culprit in my case. will get a little extra-fan to screw on top of the heatsink and see.
about the crossed device mapping i´m still clueless - but by now i don´t think it has anything to do with freenas. before erasing the bios i played around with it, and you can use it to identify drives/slots. the mapping here is equally crossed as it is when freenas is running. it said (for example) disk X in slot 0, and selecting it would light up physical slot 0. disk X however sat in physical slot 3.
the bmc (lame pseudo-ilo "LO100") is flashed to latest version i could find, 4.26. this is a used system and i don´t have any warranty-thingy with hpe - so no access to a lot of their fancy update disks. though i doubt it´d do me any good.
was looking about for information on the backplane (8 lff, 2 sas cables), but it would assume that as it has enough sas for all its disks, this is not some kind of flashable smart expander thing, but more of a dumb adapter between those big sas cables and the disks. asides from the sas cables it gets a power-cable and a small (i2c?) red-yellow-black cable to the mainboard. but seeing as the P410 card had the slots/leds/disks matching each other, i don´t think there is anything wrong with it.
i still hope someone around here knows whats going on or has had crossed mapping the issue - even though it seems to be above OS level.

 

[ninnghazad]

I came across this:
https://forums.servethehome.com/index.php?threads/hp-9217-4i4e-ports-swapped-around.12191/
Which explains how this is a whole thing about firmware, hp vs lsi, when the card was made and so on...
It seems flashing the publicly available p20 lsi firmwares just messes up the port/slot order. it is mentioned that there exists a correct firmware for this revision of the sas2308_2 D1 R5 / 9207-8i, it´s just not public.

so - does anybody have a dump of a not-messed-up p20 firmware for this card?

sas2ircu:

Code:

		 Adapter	  Vendor  Device					   SubSys  SubSys 
Index	Type		  ID	  ID	Pci Address		  Ven ID  Dev ID 
-----  ------------  ------  ------  -----------------	------  ------ 
  0	 SAS2308_2	 1000h	87h   00h:06h:00h:00h	  1590h   0041h

sas2flash:

Code:

LSI Corporation SAS2 Flash Utility
Version 20.00.00.00 (2014.09.18) 
Copyright (c) 2008-2014 LSI Corporation. All rights reserved 

		Adapter Selected is a LSI SAS: SAS2308_2(D1) 

		Controller Number			  : 0
		Controller					 : SAS2308_2(D1) 
		PCI Address					: 00:06:00:00
		SAS Address					: 500605b-0-09ce-fd40
		NVDATA Version (Default)	   : 14.01.00.06
		NVDATA Version (Persistent)	: 14.01.00.06
		Firmware Product ID			: 0x2214 (IT)
		Firmware Version			   : 20.00.07.00
		NVDATA Vendor				  : LSI
		NVDATA Product ID			  : SAS9207-8e
		BIOS Version				   : N/A
		UEFI BSD Version			   : N/A
		FCODE Version				  : N/A
		Board Name					 : H220
		Board Assembly				 : H3-25278-05D
		Board Tracer Number			: SV44738176

		Finished Processing Commands Successfully.
		Exiting SAS2Flash.

(it says 9207-8e, but it is 9207-8i)

 

[ninnghazad]

So far i have been able to find the following p20-7 firmwares for the 2308 / 9207:
broadcom p20-7 for 9207-8i
broadcom p20-7 for 9207-8e
supermicro p20-7 for 2308
all of which have a mismatch in nvdata subsystem id, but can be flashed using p14 sas2flash.
all of them also have messed up port/slot/disk orders.
the broadcom 8e version has the crossed mapping mentioned above.
the supermicro one and broadcom´s 8i version have different mapping, still wrong. slots are numbered 28-31 and the 4 disks in the second box get mapped to the same slots. also those slots are ordered in reverse. however the activity leds match the physical disks with these. the uid leds do too, but you can only activate the lights on the first 4 disks, the next 4 disks which will use the same slot numbers as the first four can not be located using sas2ircu.

i will try to build zfs using the lsi 8i firmware and keep my backups around. as long as the activity ids match i can at least go by those to identify drives.

 

[Ericloewe]

the uid leds do too, but you can only activate the lights on the first 4 disks, the next 4 disks which will use the same slot numbers as the first four can not be located using sas2ircu.

Try sesutil.

 

[ninnghazad]

Thanks for your input.
Sadly there is no /dev/ses* device.

[root@vault ~]# sesutil status
sesutil: No SES devices found

but i did not know about sesutil. however i would assume if my system supported it, the ses-driver-thing should be loaded. but i will look into it some more.