clevoir Posted December 28, 2016 Share Posted December 28, 2016 Let me apologise first for the cardinal sin of not having a diagnostics file I am running 6.2.4 on a ASUS M5A99X EVO motherboard with 3 AOC-SASLP-MV8 cards (14 data disks , 2 parity disks and 2 Cache disks (in BTRFS Raid1)), which have never stuck me as 100% reliable, I have had a number of disks go missing over time, despite cables proving OK. And also I am getting slow 60MBs parity checks too. My plan was to purchase 2 DELL PERC H310 cards to replace 2 AOC-SALP-MV8 cards, leaving just the cache disks on 1 AOC-SASLP-MV8 card for the time being. I am looking move the cache disks to 2 motherboard ports that support 6Gps. Also my motherboard has only 2 x16 slots (which work at x8), the third x 16 slot is only x4 I flashed the DELL cards using the procedure here https://techmattr.wordpress.com/2016/04/11/updated-sas-hba-crossflashing-or-flashing-to-it-mode-dell-perc-h200-and-h310/ using the UEFI procedure (unfortunately I don't have an old fashioned BIOS motherboard) I have had a huge amount of problems trying to flash both DELL and LSI firmware. After wiping the old firmware using megarec.exe -cleanflash 0, there were problems writing DELL and LSI firmware using sas2flash.efi -o -f 6GBPSAS.FW and sas2flash.efi -o -f 2118it.bin sometimes with messages saying that the adapter had not reset. For both cards I tried to install firmware, only for it to fail, then wiping the card again and then trying to load firmware again, eventually to managed to follow the procedure and get both cards running LSI firmware successfully. I thereafter replaced 2 AOC-SASLP-MV8 cards with new DELL cards using x8 slots, leaving 1 Supermicro card in the x4 slot. The array booted but 1 disk was missing, and on clicking stop a number of times I was not able to stop the array. I saw a message stating the disks were being unmounted then it quickly changed to array started again. Furthermore after 5 mins the array would lock up with no access to the GUI or server shares, also network connectivity stopped with the LED on the network interface card not being lit. Just to complicate matters I also replaced all 5 SFF8087 to SATA forward cables at the same time. However the array is working fine using the old cards and new cables, so I assume that the cables are OK I looked at the syslog and could not see any items in red or any immediately obvious problems. I've done a manual hard reboot twice now, and checked all cables and connections to the cards and 5x3 cages too. Because of the hard reset, the array wants to carry out a parity check too, but the array was showing the same problems In desperation I have removed the DELL cards for the time being, and have put back in the Supermicro cards. The array is now behaving normally, I can see all disks and a parity check is now running. Clearly it looks like the DELL cards are at fault, but don't know the next course of action and would welcome any advice. As I have now flashed LSI firmware over DELL firmware, can I wipe the card clean again and just load DELL firmware instead? Quote Link to comment
digitalfixer Posted December 28, 2016 Share Posted December 28, 2016 I'm using 2 H310's flashed to P19 with 6.2.4 without any problems. Kevin. Quote Link to comment
clevoir Posted December 28, 2016 Author Share Posted December 28, 2016 Thanks for the reply, after looking through threads on this site, it appears that people are having similar problems with P20 I'm now looking to go to either P19 or the latest DELL firmware instead Quote Link to comment
CHBMB Posted December 29, 2016 Share Posted December 29, 2016 I use P19 without issue Sent from my LG-H815 using Tapatalk Quote Link to comment
clevoir Posted December 29, 2016 Author Share Posted December 29, 2016 Thanks for your reply too. I have just wiped both cards and flashed them to DELL firmware as per OpenToe's thread here http://lime-technology.com/forum/index.php?topic=12767.825 I'll see if that works (Opentoe did not find any great difference between DELL and LSI firmware), if not I'll try flashing with P19 instead Quote Link to comment
clevoir Posted December 29, 2016 Author Share Posted December 29, 2016 Rebooted after installing 2 DELL PERC H310 cards, and UnRaid hung booting up at the CPU load command Rebooted again and the array came up with no disks missing I started the array and all was OK for 5 mins, then the array locked up with no GUI access / Network card LED had gone out and so no share access either. I had a keyboard plugged into the array, however I was not able to type any commands I am tempted to try P19 instead now, but as I had to do a hard power down, I have had to put the old Supermicro cards in back again as the array wants to carry out a parity check. Please find attached diagnostics file, I would be grateful if someone can advise me what to do next. Could this be a PSU issue? tower-diagnostics-20161229-1438.zip Quote Link to comment
JorgeB Posted December 29, 2016 Share Posted December 29, 2016 Maybe try this: http://lime-technology.com/forum/index.php?topic=29236.0 Some boards don't work well with the H310 with those pins uncovered. Quote Link to comment
clevoir Posted December 29, 2016 Author Share Posted December 29, 2016 Thanks for your reply, I was just thinking that I didn't cover some of the contacts on the card, and if this was the problem. Once my array has finshed it's parity check in 12 hours :'( I'll give it a go Quote Link to comment
CHBMB Posted December 29, 2016 Share Posted December 29, 2016 That was going to be my other suggestion but I (maybe mistakenly) thought that stopped the machine from booting, still it can't hurt. For the record, my H310 has been rock solid on P19. Also, again I may be wrong, but was version P20 the version that saw two separate releases but with no change in the version number? A silly way to do things in my eyes, but also worth looking at. Sent from my LG-H815 using Tapatalk Quote Link to comment
JorgeB Posted December 29, 2016 Share Posted December 29, 2016 Without masking the pins some boards don't boot, others have issues with one of the memory channels, sometimes not detecting half the installed RAM, other times crashing the system. Quote Link to comment
CHBMB Posted December 29, 2016 Share Posted December 29, 2016 Johnnie, am I correct in the P20 version issue or am I imagining it? Sent from my LG-H815 using Tapatalk Quote Link to comment
clevoir Posted December 29, 2016 Author Share Posted December 29, 2016 I flashed P20 first of alll and can see that some people are having problems with it. After P20 I thought about just using DELL's firmware instead, as Opentoe has it working OK As my server boots OK, I didn't even think about the covering certain contacts with tape, however I'll give it a go. I haven't seen any memory problems being flagged up in BIOS, or had any motherboard beeps codes either to incicate a memory problem. Quote Link to comment
JorgeB Posted December 29, 2016 Share Posted December 29, 2016 IIRC the first p20 had issues 20.00.00.00, there was another release after that and the latest release is 20.00.07.00 Firmware, AFAIK there are no issues with the latest and unRAID. Quote Link to comment
BRiT Posted December 29, 2016 Share Posted December 29, 2016 I'm running 20.00.07.00 firmware on my DELL H310 cross-flashed without any issues on 6.2.4 unraid on the following hardware setup: ASRock EPC602-4L, Dual E5-2670 CPUs, 128GB ECC, 5 4TB HGST drives. Quote Link to comment
clevoir Posted December 30, 2016 Author Share Posted December 30, 2016 Just rebooted after applying tape to certain contacts in the cards as suggested, see attached diagnostics file The first time I booted UnRaid hung on startup showing the BTRFS commands Rebooted again OK, but waited half an hour before starting the array. The array & network held up, and the green LEDS on the cards were flashing 5 minutes ago I started the array and it is holding up, but don't know why it UnRaid hung before? tower-diagnostics-20161230-1723.zip Quote Link to comment
clevoir Posted December 30, 2016 Author Share Posted December 30, 2016 10 minutes later and the array crashes again with the normal symptoms Network card and DELL card LEDS stop flashing, and no GUI / network access I have hard shut down, put the old cards back in and look forward to another 12 hour parity check I'm at a total loss as what to do next, I could flash to P19 but I can't see it making any difference if DELL firmware doesn't work? Quote Link to comment
RobJ Posted December 31, 2016 Share Posted December 31, 2016 You haven't actually said exactly which firmware you are running. If you are using the bad p20, then you would expect problems. You need the good releases before or after that, as johnnie.black listed. Quote Link to comment
BRiT Posted December 31, 2016 Share Posted December 31, 2016 You haven't actually said exactly which firmware you are running. If you are using the bad p20, then you would expect problems. You need the good releases before or after that, as johnnie.black listed. From his Diagnostics, he's running the following firmware, 07.15.08.00, which doesn't seem to be P19 or P20. Its as if he didn't complete the cross-flashing steps at all. Dec 30 17:09:37 Tower kernel: mpt2sas_cm0: LSISAS2008: FWVersion(07.15.08.00), ChipRevision(0x03), BiosVersion(00.00.00.00) For the record, this is what it should look like on the good P20 firmware: Nov 7 19:39:27 Tower kernel: mpt2sas_cm0: LSISAS2008: FWVersion(20.00.07.00), ChipRevision(0x03), BiosVersion(00.00.00.00) Quote Link to comment
clevoir Posted December 31, 2016 Author Share Posted December 31, 2016 Thanks for your replies, I initally followed the instructions at https://techmattr.wordpress.com/2016/04/11/updated-sas-hba-crossflashing-or-flashing-to-it-mode-dell-perc-h200-and-h310/ H310 UEFI: Run the following commands in the DOS prompt megacli.exe -AdpAllInfo -aAll -page 20 Write down SAS address. megarec.exe -writesbr 0 sbrempty.bin megarec.exe -cleanflash 0 Reboot into UEFI shell and run the following commands: sas2flash.efi -o -f 6GBPSAS.FW Reboot and return to UEFI shell. sas2flash.efi -o -f 2118it.bin sas2flash_p19.efi -o -sasadd 500xxxxxxxxxxxxx This I understand flashed the cards to P20, and I started having problems which I assumed were due to the firmware Thereafter I run the following commands using the same USB stick, but using the latest version of the DELL firmware. I read on another guide that using these commands the SAS card addresses are actually preserved, and so the SAS address does not need to be reprogrammed. Unfortunatley I don't have an old BIOS motherboard as I understand the flashing proceedure woukd be easier without having to use UEFI shell megarec.exe -writesbr 0 sbrempty.bin megarec.exe -cleanflash 0 Reboot into UEFI shell and run the following commands: sas2flash.efi -o -f 6GBPSAS.FW My first diagnostics file was taken at this point, the second file was taken after I had taped over certain contacts on the cards Looking at the file dates of the 2118it.bin file using the first proceedure and those of the same file from the latest version (20-20.00.07.00) which I have downloaded from LSI's website https://www.broadcom.com/products/storage/host-bus-adapters/sas-9211-8i#overview I have found them to be the same (11/03/16). As such I assume that even with the latest version I am still having problems. Would you suggest trying the P19 version of 2118it.bin instead? Quote Link to comment
CHBMB Posted December 31, 2016 Share Posted December 31, 2016 But in response to BRiT's query, what does it say now? Still the same thing? Quote Link to comment
JorgeB Posted December 31, 2016 Share Posted December 31, 2016 Firmware 07.15.08.00 is Dell's firmware, both that and the one from LSI (except maybe the first p20 release) work with unRAID, if you have issues with either I'm guessing a bad controller or some incompatibility with your board, can you test it on another PC using windows or an unRAID trial key? Quote Link to comment
clevoir Posted December 31, 2016 Author Share Posted December 31, 2016 In response to CHBMB the cards are currently wih DELL Firmware 07.15.08.00 installed. I have had both cards running another PC OK and they show up in Windows device manager as DELL H310 As suggested it looks like a problem with the cards (however they booted up fine showing their BIOS page installed in a Windows PC before I flashed them) or an incompatibility with my motherboard Quote Link to comment
CHBMB Posted December 31, 2016 Share Posted December 31, 2016 In response to CHBMB the cards are currently wih DELL Firmware 07.15.08.00 installed. I have had both cards running another PC OK and they show up in Windows device manager as DELL H310 My motherboard has 3 x 16 physical slots 1. Graphics slot x16 2. x16 (electrically x8) and 3. x16 (electrically x4) When cards are installed in slots 1. and 2. both run at x8 - which is the case when the DELL cards are installed I understand that some motherboards have problems running any other cards in the graphics card x 16, and wonder if this could be the problem (although a AOC_SASLP-MV8 card works OK but this is only x4)? For info the network card which stops working is a seperate card installed into a PCIe slot, so could cards be affecting this too So, is there no way you could access a PC to flash them correctly? Quote Link to comment
clevoir Posted December 31, 2016 Author Share Posted December 31, 2016 I first flashed them to P20 using the PC which I use as my UnRaid server After that I then flashed them to DELL's firmware using another PC Both are UEFI as I don't have an older BIOS PC I had problems using the first PC but flashing using the second PC was fine, and I was notified that the firmware had been installed correctly Quote Link to comment
JorgeB Posted December 31, 2016 Share Posted December 31, 2016 Forgot you're using two controllers, so it's not likely a bad controller, my vote is for a board incompatibility, did you try just using one at a time? You could also try them them on another PC with an unRAID trial key. Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.