t33j4y

So, I had this happen to me as well. Due to an update of cache pool drive, I actually have very recent backup version od the cache drive version with the docker (SabNzbd) that disappeared after updating. Can I somehow roll that back?

Alright, makes a lot of sense :-) Does the HDD temp sensor work off of the highest reading across the disk pool or the average of the pool?

You may have to factory reset the bmc. Also remove the .config files from the ipmi plugin directory on the flash drive. And make sure that the load config toggle switch is off on the sensor editor. Found this and toggled the switch to "OFF", after restarting the server. So far it seems to be holding the config without going full blast. Can see on the Readings tab that RPM is actually adjusting. *fingers crossed* Still wondering why the CPU fan - a Noctua NH-L9x65 rated at RPM 600 - 2500 ± 10% seems to never go under 1000RPM, even though the fan speed minimum is 31. So compensating for the ± 10%, 31% gives a range of 775-852 for lowest RPM, as long as temp is below 40 degrees. But even when temp is below 40 degrees, it still hovers at around 1000.

You asked about the error messages - here's an example: So, some progress now. Setting to Full Speed seems to have helped - pulsing is gone and the plugin seems to have regained control over rpm. I have edited the three Lower Critical values for the case fans to be at 300/300/300 and the cpu fan at 500/500/500. Case fan thresholds for temp are 35/40 and for rpm 40/100 CPU fan thresholds for temp are 40/60 and for rpm 31/100 I'm gonna fiddle around with it now to find final and sensible settings :-) Thank you so much for putting me back on track! Edit: Spoke to soon :-( After changing the RPM settings, it is now back to full blast... Did the following to recover: 1. Uninstalled IPMI plugin 2. Removed IPMI folder in /config/plugins/ 3. Shutdown server 4. Reset BMC to factory (from within BMC) 5. Changed fan speed to Full Speed 6. Started unRaid 7. Installed IPMI plugin 8. Changed Lower Critical Values: case fans to be at 300/300/300 and the cpu fan at 500/500/500 and set settings to reload on unRaid start 9. Set IPMI plugin to - Case fan thresholds for temp are 35/40 and for rpm 40/100 - CPU fan thresholds for temp are 40/60 and for rpm 31/100 10. Enabled Fan Control Still, full blast :-/ I should have just kept my fingers off of it, instead of trying to adjust more. 11. Repeat the above, plus full power down and mains lead removed. And now, it worked. But only for 2 mins after which it stepped RPM up to full again on fans 1-4. Even if the temp thresholds weren't reached :-( CPU fan is at 1000 which does not correspond to the threshold temp of 40 as it is at 35.

Thanks! I will try out what you advise as soon as I return from work. I can definitely run the commands manually if you want. What are they, please? :-)

So, I am at my wits end here :-( Trying to get this plugin working on 6.4.1 with 4x Noctua PWM case fans (2x 1200RPM and 2x 1500RPM) + 1x Noctua PWM CPU cooler (max 2200RPM) on a SuperMicro X11SSM with support for PWM. I set the SuperMicro fan control at "Standard" in BMC and install the plugin. Configure case fans to react to HDD temp and CPU cooler to react to CPU temp. Configure fan control for case fans to min/max temp of 30/40 on HDD and fan speed to 40%/100% Configure fan control for CPU fan to min/max temp of 30/60 on CPU and fan speed to 40%/100% If I leave it like this, the fans pulse up and down endlessly, throwing hundreds and hundreds of warnings in the event log. I then tried adjusting the lower thresholds as advised earlier in the thread to 300/300/300, the idea being that if the plugin is set for lowest fan speed above this, it will never trigger bios to try and adjust (causing the pulsing) and everything is good. Unfortunately, what happens is that all fans immediately go to full blast. This change seems irreversible, at least through "conventional" ways. Tried disabling plugin, adjusting lower thresholds back to default, uninstalling the plugin, cycling through the modes in the SM IPMI interface. The only thing that reverts it is resetting BMC to factory default. Any advice on what to do would be much appreciated. Thanks in advance!

That sounds quite straightforward - will do that! Thanks :-)

I'm running an Ark server in a Windows 10 VM, but for various reasons I'd like to set this up on a dedicated PC. Is there a guide on how this is accomplished in the easiest way - or was setting it up as VM a one-way street? :-/ Thanks in advance for any insights, links and help! :-)

Would like to buy an i5-3470T for a low power build. Am based in Denmark.

Thanks :-) Disks are on a cross flashed Dell H310 so I can only swap fwd breakout cable. Since there are no errors on remaining 3 disks on same channel, can I safely rule out that it is not the HBA acting up? Regarding cooling, I am actually trying to move disks around to decrease temps - the chassis is a Node 804 - so there is not hundreds of options. Maybe I should get that custom chassis done...

Ok, updated diagnostics here. tower-diagnostics-20180202-2334.zip

Would that not be because the system disabled it?

Diagnostics here. tower-diagnostics-20180202-2108.zip

At work now, but will do when I return tonight. Thanks :-)

Woke up to a succesful parity check according to notification, but one of my parity disks have been disabled :-/ What would the best course of action be to resolve? Thanks, Thomas syslog.txt

Thanks - will look into this. It seems that there are some improvements in 6.4 regarding certificates and encryption that I should probably understand before diving in. Thanks again for the swift response!

Oh, wow. I was fully unaware that the disk was "special" - I thought it was just a regular 6TB drive :-( The disk is run from a DELL H310 crossflashed to IT mode. Will this be a problem for the controller or is the fact that it is "recognized" proof that it will work, once I use 6.4? And how do I go about updating to 6.4 now that the array is degraded? Should I go back to the old 4TB drive, upgrade and then replace to 6TB?

I have come across a strange problem that I would like to request some input on. One of my disks (4TB) was showing signs of fatigue (SMART error 187) so I prepped a new one (6TB) and precleared it. It's disk5, on sdl. Followed the instructions on how to replace: 1. Stop the array 2. Unassign the old drive if still assigned (to unassign, set it to No Device) 3. Power down 4. [ Optional ] Pull the old drive (you may want to leave it installed for Preclearing or testing) 5. Install the new drive 6. Power on 7. Assign the new drive in the slot of the old drive 8. Go to the Main -> Array Operation section 9. Put a check in the Yes, I'm sure checkbox (next to the information indicating the drive will be rebuilt), and click the Start button 10. The rebuild will begin, with hefty disk activity on all drives, lots of writes on the new drive and lots of reads on all other drives But I get stuck at step 10 as the rebuild doesn't begin at all. GUI becomes unresponsive for some time and then returns without having initiated any rebuild. I can start the array if I set the disk slot to no device. Disk is recognized and shows up with a blue icon for "new drive". Really don't know what to do. Have attached the preclear report and syslog. Hope someone has some words of enlightenment :-) preclear_report_S4D0P7DJ_2017.12.24_23.01.03.txt tower-syslog-20171228-1910.zip

Can't really speak for the stock fan as I didn't use it on the e3-1230v6 I just set up. I went for the Noctua NH-L9x65 which keeps it cool and quiet. If you go for this, make sure you get the socket 115x version as there is a pure AMD version as well.

So, some days later and a lot of both grief and relief has passed. This is gonna be longish but might hold some interesting lessons for others with the X11SSM. BMC and BIOS are at 1.37 and 2.0c (board was delivered with 2.0b, so I upgraded). H310 has been modded with electrical tape to cover B5 and B6. 1. H310 HBA cross flashing is in itself not troublesome. Creating the right bootable USB for it , is. I could not run the flash tool executables without "misfiring" in weird error messages like stuff about mouse driver not being loaded (why the hell this would be important for a DOS shell, running CLI programs is beyond me), memory not available etc. Finally got it working using HP USB Disk Storage Format Tool and Win98 files downloaded from allbootdisks.com and succesfully cross flashed the H310. I feel kind of lucky that I succeeded because when I rebooted using the same USB boot key half a day later, I could not get any of the tools to start (I was back to the "mouse driver missing" type of errors). Seems very hit or miss (and also made me realize how much I had forgotten about DOS config.sys of olden days - which made me sad, but that's another story :-) ) 2. Because I wanted to make absolutely sure that everything worked as expected, I installed a trial version of unRaid on a different USB key and set mobo to boot legacy USB key only. and plugged in two extra disks I had lying around. Both on mobo controller. Booted and let unRaid start in non-GUI mode. Accessed via web interface, formatted disks, created share, put files on it. Powered down, moved data disk to H310, restarted. Everything went swimmingly, disk was recognized, files were present. Great stuff. 3. I hooked everything up - cache disks on mobo controller and the rest on the H310. Plugged in my USB key with licensed unRaid. Booted into unRaid non-GUI mode. Did not get an IP for unRaid (this may be due to another issue - see note below). Strange, but it seemed like there was no link at all. 4. Rebooted to try and acquire DHCP. Now unRaid would not boot and got stuck in an endless loop where bzroot and bzimage would load, and then board reset. Ok, back to last known good config - the test setup. unhooked all disks from mobo and mini-SAS connectors from H310. Swapped to test USB key. Reboot. Same thing. Loads bzimage and bzroot, then resets. But this setup just worked an hour ago! 5. Searched on forums and found other X11SSM owners who cannot boot into unRaid non-GUI mode either. They can boot a plethora of other stuff (including Linux distros) without issues at all, but not unRaid. But if they boot to unRaid GUI instead of unRaid non-GUI, it works. No one knows why. It's not even a problem on all X11SSM boards - only some. And some have reported that if they load optimized defaults for BIOS and touch nothing in the BIOS setup, they can boot unRaid non-GUI (resetting to optimized defaults didn't work for me though). So, I tried booting into unRaid GUI and lo and behold, it worked like a charm! 6. Ok, so back to hooking up all disks and the licensed unRaid USB key. Boot in unRaid GUI, change config.sys so that unRaid GUI is now the default (for unattended reboots) and everything worked as expected. Disk positions were intact from my original C2750D4I. Am now running parity check at > 150MB/s which is a 40% improvement from my old server :-) I would still like to boot into unRaid non-GUI instead of GUI (personal preference) but can live with this. Note: It may be that one of the network cables is iffy and that between the stage of "test setup" (bullet 2) and then "hooking everything up" (bullet 3) I might have swapped the iffy one to NIC #0 thereby not getting an IP address. However, since both network cables were inserted, I would have thought that unRaid would fall back to NIC #1. When I swapped them, and did not put a cable in NIC#1, I got an IP address. So I need to figure out if I have an iffy cable or if the board does not like to have two cables attached. But I have a feeling that there is more quirkiness to this board that I have yet to see.

Yes, they do not have RAID capability but are "dumb" Host Bus Adapters (HBA).

Thanks to both of you! My C2750D4I isn't working at all, but my array never autostarts anyway. I've had good response from SuperMicro in the Netherlands and have been asked to RMA the board. Hopefully, they'll replace it and I can cover some of the expenses for the new load of kit by selling it off once I get the RMA back.

My existing unRaid server mobo stopped responding a week ago and would not post anymore. Could speak to BMC I have since been getting new parts home and am planning to migrate during the weekend. I am unsure how painless this is going to be and am asking for some advice to avoid problems with disk assignment. My USB stick is intact Old setup AsRock C2750D4I - with 12 SATA ports (split across an Intel controller and a Marvell controller) 2x 8GB ECC DDR3 RAM 2 parity drives 2 cache pool SSDs 6 data drives (all XFS) New setup SuperMicro X11SSM w/ Xeon E3-1230v6 - with 8 onboard ports (Intel controller) and 8 ports on a DELL H310 controller card which will be crossflashed (and I will use the tape mod) 2x 8GB ECC DDR4 RAM All drives should be migrated My worry is that although Unraid 6+ has become clever in terms of recognizing drives and placement, I have read several times that a controller card can report disk ID in a different format than onboard controllers, which might cause trouble for me. Should I be worried? How do I avoid creating trouble for myself? Or will everything be ok? Thanks in advance for thoughts, guidance and suggestions!

Adrian, if you go for rebranded LSI controllers, be sure to check the advice given to me by johnnie.black in this thread: It only pertains to some controllers, but I am unsure what you're ending up with, so just in case :-)

Wasn't aware of that - seems straightforward though. http://yannickdekoeijer.blogspot.co.uk/2012/04/modding-dell-perc-6-sas-raidcontroller.html?m=1 Thanks again!