Onemorethan0

Been using unraid about a year now, had an old 3800X gaming machine with the stereotypical 400$ board for no real reason but it had 10 sata ports and so I setup my array like that. Well flash forward some time and through some unfortunate events that board died ungracefully. I was stuck with an extra board I had with only 6 ports. Welcome to the PCI-E Sata adapter days... obviously unacceptable I went to work determining a cheap and thrifty way around this. (while addressing 6 drives each would run at 16mb/s during parity check, by the end of the story im running parity checks at 220mb/s on all disks) So I picked up these 2 old CM240 M3s 256gb of ram each, 2X12c/24t @ 2.7ghz E5 2560-v2 or whatever he even threw in 24 SAS 2.5s averaging 600gb - 1.2tb in size 10k rpm approx 30k hrs runtime on most. They also had intel 350 4 port nics and well the title 9271CV-8i. Immediately I'm like this is useless... 2208 based trash. Then I see about cross flashing... I spent this weekend trying like heck to crossflash and what i found is that when it thinks its a 9207 it works but the moment it touches my sas expander it is like nope I go no further. I then took the unflashed one and tried JBOD and while it was tedium to setup and would be tedium to continue with, it did work as intended I was getting smart data and total throughput was maxing out around 3.3gbps and it was the drives at 100% controller seemed to have plenty of legs. I turned to the other card to try and flash it back. This is where things get very strange and I can't explain what's going on and maybe some one has a guess. I flashed 2208_8.rom the old one for 9207. then sas2flash'd the 9207-8.bin then to reverse I cleared it reflashed with 2208_16. I did use megacli to enable JBOD but I did not set any ports or drives to be JBOD. I put it in my unraid server boot it and.... It just works???? I see the drives they pass smart data and for all intents and purposes unless I look at the actual controller its connected to in system devices I couldn't tell it wasn't on the on-board sata. Is this just my misunderstanding of JBOD and that because the OS is allowed to see it its also allowed to just make the call and use it as a normal disk and I didnt have to manually set 24 drives to jbod on the other server or what? This is in no way a post suggesting you or anyone else should buy a 9271cv-8i I whole heartedly believe good advice when its given and that using something with no additional cache and battery is going to make life easier and more reliable, probably faster too. I just want to understand what my second 9271 is doing here, if jbods enabled but i didn't set any drives to be jbod and not only that but my drives were on a mix of onboard sata and pcie add in card sata and yet all immediately recognized with the same name on boot and the array was even started when the machine came up. How's that even possible?

Cable swaps didn't seem to change anything, I guess where I'm at is going to have to be good enough for governement. Swaping cables did bring a couple back enough to get the array up and running but that spin cache is dead for good as best I can tell at this point, I'm going to wait till i replace the board so I can unite them all on the same sata controller to see if it makes a difference but I am not optimistic. Thank you for the assistance.

Yeah this is all exceptionally temporary to pull off wedding photos shot at previous engagement to send out to client. So if I can get it stood for a couple hours that would work for me. Thank you for your insight this alligns with what i've been seeing as i've replaced most of the sata cables at this point seemingly when the PCH went it surged the entire system when it cooked the 2600X but all not bad since I had a 3800X to replace it with soooo in the end it wil lbe a benefit, I just need a new nice AM4 board to replace my cheapy backup. Wonderful. I have a direction at least. Will report back with results of replacing sata cables and if that doesn't work, I have one last drastic measure which is to basically transplant all drives to my main gaming machine as it has a crosshair VII hero wifi board that has more than enough sata support to make it happen.

My apologies, been a minute since I was looking for assistance. Here you are.arkenstone-diagnostics-20231019-1300.z Thank you!

So as the title suggests, I had the most catastrophic psu/board failure that it blew the PCH and I killed a graphics card, then sacrificed a second to the blood gods just trying to troubleshoot. So I slapped a new proc and board together and unfortuantely it has a far less capable sata controller went from 10 to 6... Then found out 2 don't work with an m.2 present. So I threw in a slow boy sata expansion card. I thought this was the issue but apparently not since when its on the actual board sata ports it does the same thing. So this is the unusual situation behind this drive, It failed once before, I put anotehr one in and kinda botched it so it wound up that it declares its RYW which actually found later it works so its in the system and it says nah, wrong drive, the drive I turned into it is the drive that matches that 2E4166 which is a WD 4tb drive. Its available as well. Selecting either the original or the replacement that took its place and was working when it went down, it still tells me I have the wrong disk. Is there some way to force it to try with the drive of my choice or am I screwed? luckily this is the cache ingestion pool sooooo no REAL major loss... I think.... If i can't recover it but boy would I like to. Many thanks for any opinoins.

One last final update for anyone wondering, long story short, the 980GTX the faithful of my twin sisters that got me my first working 4k setup with SLI 980s has failed. Removing it from my unraid server got me truly stable, i've now transferred from a trialer to a customer and am happy thus far with the purchase. Great minds here, I was able to solve a plethora of issues I ran into along the way chaining dockers together and managing permissions just by searching the forum and not until this whole debacle did I need to post. Thanks for the help.

Marking as solved for now. I was able to get the system running again, parity check completed and repeating the same stress tests that would crash it before it has remained stable. Will update if it changes. Thank you for the help.

Thank goodness for overly expensive boards. Dual bios ftw. so bios 1 is corrupt on my board now... not 100% sure how that came to be but i set my clock settings back in line with your guidance as well as disabling c-steps completely in bios. We shall see how it goes if it lasts 28 hours id say we are good.

Thanks for the response. I went to go change the memory settings and now the systems down for the count with a P0 on the debug that doesnt match any code listed for this board oh yay. So effectively im guessing either the PSU ate it or the ram burned up lol

So had unraid about a week now or so, everyday around 5am it crashes. Not sleeping, nothing I can determine in this log. The only common theme is it will have recently changed the fan speed and then boom logging a new start up. All of the syslog I've gathered so far from mirroring the log to the flash is attached. Any further insight into these crashes would be helpful as I'm kinda past the point of return as I have so much data on here it would be a true pita to pull it back out and move away from unraid. I'm hoping its just some obviously noob issue. My big fear is its because this runs on an older 1800X amd cpu and AX370 gaming 5 board. syslogcrashes.txt