matthewmalek

I'm having issues getting the ThunderboltEX 4 and mobo to play nicely on a bare metal Win 10 installation, so I think I'm going to call it quits. I'm going to return some of my recent hardware purchases and instead build two separate systems.

Update: I now believe the IOMMU configuration above is OK. The NHI is what a VM is going to talk to, and the system will use PCIe over USB 4 to implement Thunderbolt (hence why it's a USB controller). I made sure PCIe tunneling over USB4 was enabled in my BIOS (it was by default), I attached the two Thunderbolt devices to my Windows 11 VM, and booted it up. Note that when booting with the IOMMU groups bound to vfio, there is no longer a 'security' file for the device because it is no longer being mounted by unraid. I had one unrecognized device in Windows so I installed this Intel Thunderbolt software (https://www.intel.com/content/www/us/en/download/19402/thunderbolt-3-and-4-dch-driver-for-windows-10-for-intel-nuc.html) which has an overview page (attached). It seems to think that PCIe tunneling is disabled and the tooltip says it can be enabled in the BIOS. Why would both Unraid and the Windows VM think that PCIe tunneling is disabled and how can I troubleshoot this further? Thanks! multivac-diagnostics-20220601-1847.zip

Hi all, I have an ASUS motherboard with a Thunderbolt 4 header: https://www.newegg.com/p/N82E16813119540?Item=N82E16813119540 as well as this Thunderbolt 4 PCI expansion card, the ThunderboltEX 4: https://www.newegg.com/asus-90mc09p0-m0aay0-pci-express-to-thunderbolt/p/N82E16815293048?Description=thunderboltex 4&cm_re=thunderboltex_4-_-15-293-048-_-Product&quicklink=true running under Unraid 6.10.2 . The motherboard bios has been updated. I plugged the card into the first PCIe slot on the motherboard, and connected both the Thunderbolt 4 and USB headers to the mobo (I didn't connect the molex power b/c I believe it is optional for enabling quick charging Power Delivery, which I don't need). In unraid's System Devices I can see a lot of "Thunderbolt" items, but only the USB controllers have checkboxes next to them. I've tried booting with all 4 ACS modes and it always looks the same (see attached image). Why am I unable to select the PCI bridge IOMMU groups to bind to VFIO at boot? EDIT: The command cat /sys/bus/thunderbolt/devices/domain0/security outputs "nopcie". I couldn't find any Thunderbolt security settings in the motherboard bios to change this. But perhaps this is part of the reason? Thanks for the help, -- Matt

Thanks Johnnie. I haven't had time to look into it until now, but changing disks 6 and 7 to different ports fixed the issue. I will also look into changing the controller type.

Oh wow, didn't realize there was such a thing. Very handy. Attached. multivac-diagnostics-20180204-1217.zip I've also noticed that the issues appear to by with my two latest discs 6 and 7, both of which are xfs whereas the rest are reiserfs. I have tested and discovered they do not spin up anymore. I have not restarted yet, but if history holds, they will spin up on reboot and be accessible until they spin down, for good. Failed attempt to spin up up disc 6: Feb 4 12:41:23 Multivac kernel: mdcmd (52): spinup 6 Feb 4 12:41:23 Multivac kernel: md: do_drive_cmd: lock_bdev error: -2

unraid: 6.3.5 For quite a while now my system has been not listing all my files and directories after having been on for a short while. When I reboot the system everything is available, but after a while certain user shares start having many of their files go missing. This is true even when I ssh into the system and list the files directly on its filesystem. I believe the issue is between the discs and the kernel, as is made clear by the logs I've attached. But I don't know enough about what's going on to make a real diagnosis. I would like to make sure my data is safe in case one of these times I turn on my system and my files never show up again. I included a log that shows just the errors. There is a constant "Emask" kernel exception throughout, but I believe the bottom shows the real error as I try to access an affected user share. If anyone can suggest a fix for this issue, I would greatly appreciate the help! syslog_ERRORS.txt

Thanks, Frank. Very helpful post. I changed the SATA cable (seemed to be connected well on HDD side, but was a little wiggly on SATA card side) and I am currently rebuilding the disk. I will definitely look into a UPS in the future when I can fit it within budget. I think all my issues are as solved as they can be (the main issue of unraid freezing is gone), so I think this thread has run its course. Big thanks again to all who posted!

Hello again, forum goers. I'm adding to this thread again for two reasons. First, to update the status of the main issue (freezing), and second, to report a new disk issue. :'( So I think that the issue of unraid maxing out the cpu and being unresponsive has been completely fixed by fixing the clash between Plex's nighttime actions and unraid's cache mover. I changed the Plex's scheduled tasks to run between 1am and 3am, which avoids the mover's standard activity window. But now for the new issue. I previously had disk 5 become disabled, and I don't think I ever figured out why that was, but I submitted the SMART report to this thread, had it looked over, and was reassured that the disk seemed to be OK. So I cleared and rebuilt it, and things were running ok for quite a while. Then, a couple days ago as the vacuum was running, the circuit breaker flipped for our living room and cut the power to my server. Fortunately, nothing was being moved to it at the time. I turned it on and got green circles all the way down, so everything seemed fine. But today I attempted to access my music share for the first time since the power incident and it appeared totally empty (I added it and disk 6 at the same time, so all of it is likely on disk 6). I noticed that disk 6 now has a red X, and is supposedly being emulated. After a couple of attempts to access the music share again, the contents did eventually appear (presumably via emulation). I have attached the SMART report for disk 6 and I was hoping to have someone determine if the disk is actually functional, or if it needs replacing. Also, I am wondering if it it possible for what is occurring with disk 6 to have come from the unclean power down. I don't like to be rebuilding disks every month, so if there is a bigger problem at hand, I would like to figure it out Sorry for the long post, and thanks for reading. Getting my server to a stable place like it used to be would be great, and I appreciate all the advice to help get it there. WDC_WD20EARX-00PASB0_WD-WCAZAF678954-20160625-0033.txt

I took out disks 5 (dead) and 4 (OK) and plugged them into my desktop for comparison. I feel like I've done this before, and working disks show their file systems to Windows and let you explore them. Both of these came up as OK in Windows, but with a primary partition that was entirely empty. So I don't know what to make of that. I plugged them both back into the server and they show up as 5 dead and 4 OK. Although, now that I think about it, I don't know if the server actually thinks 5 is dead. It knows it's there, so I think it may have disabled the disk for another reason. I've attached the SMART report for disk5. Can someone assure me if it is OK or dead? Disks 5,6, and 7 are on the new PCI SATA expansion card that I replaced. I remember before, disk7 did not show up when rebooting from one of the freezing episodes, but when I rebooted the system again (after a proper shutdown) it was recognized and has worked since. I think that card just doesn't like improper shutdowns, and maybe unraid has marked disk5 as BAD even though it's ok. WDC_WD20EARX-00PASB0_WD-WCAZAJ896221-20160515-2343.txt

UPDATE So, I came home this morning and my server was off. Not just powered down, but the switch on the back was flipped. Apparently one of my family members turned it off last night since it was giving out single, intermittent beeps (from the speaker it seems) every 30 sec or so. Does anyone know what could cause that? And then, after powering it on, an odd squeaking noise was coming from my server (not from the internal speaker), and suddenly the web interface indicated that disk 5 of 7 is not working (red X) and is being emulated. So I guess this disk is dead... Now what I'm wondering is, is it possible that this issue and the previous issue are related? Can an failing disk cause issues during a parity check? This disk likely wasn't being used to write new data, since it's been fairly full for a while, but I assume it would of course be accessed during a parity check. (And I assume it would also be used on occasion when streaming media from it, but the freezing issue only occurred overnight.) I don't see how that can easily be related, but it seems like a big coincidence that they both occurred together, especially since I've never had a failed drive or freezing issues before. I did recently have a PCI SATA expansion card fail on me, but I replaced that a couple months ago. So I guess I'll get a new disk to replace it, and take it from there.

Wow! Sorry for not being more active in this thread. I assumed I was automatically subscribed, but it turns out I had to do that manually. I assumed no one had a response to my issue. Well, I'm back now that my system is having the same issues again. To answer a previous poster, yes, I am using Plex. And as I mentioned before, I am using a cache drive as well, and the tv show and movie shares are allowed to use it. So now I am wondering, is it possible for Plex and the cache drive to not play well together? I can't imagine I'm the only one using these two things. I know the cache drive tends to move items in the early morning, and someone mentioned that Plex does some things in the morning as well. (I adjusted the Plex schedule so it shouldn't overlap with cache mover anymore. Maybe that will fix things.)

Hi, I upgraded to V6 a few months ago, and it usually runs great, but I recently have been experiencing freeze-ups every week or so. When this happens, I cannot access the web GUI, or access the SMB shares. The server seems to continue to display video, but I do not have a keyboard attached to it, so I do not know if it would respond to keyboard inputs. I know it does not respond to a press of the power button. (I have to hold in the power button to get it to power off. I think that is a mobo action, not a sign that the OS is responding in any way.) I am using a cache drive, as well as a few plugins (Recycle Bin and others.) I also have the Plex Media Server docker installed. The way I know I have a freeze-up is not always because I cannot access my shares, but because the server gets unusually loud. Placing my hand behind the server, I can feel it is fairly warm (warmer than usual, I believe.) Therefore, I think that the CPU is stuck in some sort of cycle and it is getting very hot, forcing the fan to run high. I have searching the forums for similar problems, and while other people have freeze-ups with V6, I could not find a situation that was similar to mine. I am running 6.1.8, but just saw that there is an update for 6.1.9. I will make the update soon, but nothing in the changelog made me think that it would necessarily fix the problem I have been having. (Diagnostics attached.) multivac-diagnostics-20160428-1508.zip

I'm connecting using Windows Explorer from a Windows 7 machine to an UNRAID 4.7 server. I have "Share security" set to "User Level," as it always has been. There are two users. The 'root' user is there by default, of course, and the only other one is 'user' which I have created without a password. It doesn't seem to be the case that root always has r/w access. I have a share with the export set to read-only, and my "valid users" set to 'user' with "exceptions" and "invalid users" both blank. It will prompt me for a username and password, and entering my 'root' credentials gives me an "access is denied."

Well, now I feel stupid, but I still have a question. So I had messed around with my share permissions the other day, but I guess they didn't actually take place until I rebooted, so I didn't realize that my changes didn't work. I had my shares setup to be in read-only mode, with 'user' as a valid user, and 'root' as an exception, so as to let me read and write. But I guess that doesn't work for some reason. But I had it the other way around before (shares set to r/w with 'root' as valid and 'user' as exception) and that didn't work, either. So I guess it boils down to this: What is the correct setup in order to have a share be accessible to some as read-only, and to others as read/write? I used the wiki in order to do my original configuration, but that didn't seem to work. And according to the wiki, both methods are equally valid.

This is a fairly strange problem, but I'm hoping someone may be able to help me fix it. First, I'll explain my user accounts for unraid. I have the 'root' account which has full r/w permissions on all of my shares. The only other account is the 'user' which has read-only permissions on all shares. I leave my desktop logged in as the 'user,' but I wanted to switch to 'root' in order to add some files. I wanted to find a more convenient way to switch users than rebooting my pc, so I found that you can type "net use * /d /y" in the command prompt to get Windows to forget its credentials. After doing that, it would prompt me for my username and password when trying to access the unraid network shares, which is good, but my 'root' password wasn't working. The strange thing is, when you normally enter an incorrect password, the login window states "The specified network password is not correct." But when I would enter the correct password, it would instead say "Access is denied." So that's where I'm at. I can't add things to my unraid shares because I can't seem to login with my root user. (And, yes, I am sure it's the correct password because I used it to change around things through the web interface.) I tried making another user with the same permissions as the root user as sort of a backdoor, but I get the same "Access is denied" crap. So, if anyone knows why I'm being denied, I would be very grateful. I've restarted both machines, but the problem persists. Thanks for you help.

I did find this (http://r3dux.org/2009/07/how-to-force-windows-to-forget-a-network-share-password/) solution, but I get an error stating that an active process is accessing the device (which isn't true)... There are also ways to use the Key Manager built into Windows, but it doesn't show up there.

Have you found a way to switch between users? I don't want to leave my desktop signed into unraid with all read\write permissions enabled, but sometimes I need to move files around, of course. Is there a way to get Windows to forget its username/password for a network share so that I can sign in as a different user? I'll continue looking, but for some reason this isn't a more discussed topic. Let me know if you've discovered anything youself, aside from rebooting your machine. Thanks.

Alright, so I rebuilt the drive (again) and replaced the SATA cable. I'll keep an eye on it, and if it happens again, I'll get a different SATA PCI card. Thanks.

I attached the files mentioned. (I had to add the '.txt' extension to the syslog for it to upload, so remove it if need be.) The difference between this problem and the problem that I linked to before is that, before, my overall status was red whereas now it is orange. I wouldn't be surprised if it's the same problem as before, but it seems to be a little different. So, if this is the same problem as Joe L. has described both times, then it looks like there is a problem either with my power connection, or the Sata connection (which could be anywhere from the cable to the pci card, to the pci slot on the mobo...). I'll start by changing the cable. Is it possible that the pci card isn't being recognized properly sometimes, or is this likely a hardware error? Thanks. smart.txt syslog.txt

Great... Well, I was about to go out of town for a few days, and I was checking the array to make sure everything was working so I could take it with me. But now I'm just going to turn it off and come back to it when I return. I'll post those reports when I get back this weekend. Thanks for the quick response. Note: I've been adding things to the array, so are there things that are written to parity and not on any actual data drives? The red drive is the drive where new data would go if it was working, since it is clearly the most free, but I don't know if it would have put the data on the other drives instead, or if it would just pretend like the red drive was working and update parity as necessary.

So, attached is an image that shows the current status of my unRaid setup. It seems to be working alright, but, of course, it's a concern whenever all the status lights aren't green. What's interesting is that I've had a similar issue ( http://lime-technology.com/forum/index.php?topic=15785.msg147007 ), but a little different. The same drive is in question. This drive is the only drive that is connected to a PCI card that I added to increase my available SATA ports, so I'm wondering if that has something to do with it. Anyway, I couldn't find what the orange indication with the red drive is meant to mean, and I wonder if looking at these problems together will give better insight into both of them, so I decided to post on this forum, again. So, thanks for any help you may be able to offer. I may end up replacing the SATA cable in addition to doing any other fixes that may be suggested.

Well, whatever happened, it seems to be working fine now that it's rebuilt itself. Thanks for the help.

Alright, will do. Thanks. As far as the power goes, it makes sense that a gap in the connection to the drive would be the source of the whole mishap, but I'm not sure how it would have happened; I checked the power connection and everything looked fine. I'm wondering if it could have been caused by a bad SATA connection. Interestingly, it is the first and only drive I have attached to the SATA expansion PCI card I recently put into the server. Everything looked fine there, too, though. I have a 500W psu, which I'm assuming is enough to run the 5 drives and 2 pci cards (gigabit network card, and SATA expansion card). And now that I think about it, I remember that the last thing I moved onto the server (a video) seemed to go over just fine. I remember copying it to the server and then watching it immediately afterward. Would it be possible for it to have copied onto parity while the disk was "missing"? Either way, I'll be rebuilding the drive overnight. If everything goes fine, I'll change post to [sOLVED]. Thanks, again.

Hi, I attached a screenshot that should explain my current situation, but I'll add some context as well. I was using my server yesterday and everything was going fine. I powered it down and powered it back up today. There was nothing being shared, so I looked at the GUI and found this. I don't know why it assumed the drive has been replaced, but it is acting as if there is a different drive in there. I could just rebuild the drive, but I was wondering if there might be some underlying issue that I should take care of. Also, when I do get around to rebuilding it, is there a way to do so that won't take as long since it really is just the same drive and same information? Thanks for you help. I'm not going to touch anything until someone can shed some light. syslog.txt

Alright, perfect. Thanks guys. I believe the method of buying two new drives and building parity after the transfer is the road I'll go. It was my plan B just in case there wasn't a way to do it without so much data movement. Although creating/verifying a checksum file will actually ease my mind a little bit. Sidenote: Is there a way to allow only one pc to write to the server? I'm sharing a network with a few other people and I would like my desktop to be the only device that can add/remove files from the server. Although, I would like for everyone to be able to read from it. This is mainly as a precautionary way to keep anyone who jumps on the network from changing data on the server. Thanks again.