Dela7460

I reseated the drive and It came back to life. I'm backing up all the data now. And ordering two Nvme drives to raid them together as a new cache drive. My VM's are the only thing I'm concerned about; it seems that every few months I'm having to deal with some issue regarding my VM's

Woke up this morning to find out we had a power blip; and found all my dockers and VM's missing. Unraid on UPC and I didn't see any messages or alerts about a reboot. Either way when I navigate to cache drive it is completely empty no folders no nothing. I can see on the dash board that there is usage. I restarted the server and now I find that Cache drive is being undetected; my only guess is that the cache drive died without warning. I had backups running at one point so I hope that still intact. Has anyone experience a cache drive die on them that was similar to this. 1tb SSD it was a Crucial ct1000mx500ssd1 drive May 17,2022. There was a warning message on be according to the forums it was ok to ignore; I can't seem to find what it was. If I do I'll post it. tower-diagnostics-20240204-1243.zip

So it looks like all my troubles were caused by a bad brand new NVMe drive I got from Amazon. creating a pool of cache drives with it caused all kinds of madness to happen. Probably why it was saying it was read only and in other places it would save device 2 was missing. I would recommend people stay away from fanxiang branded drives yes they are super cheap but seem to be total crap. Waiting on Parity to finish to restore docker and vm's Thankfully I did have backups of all that.

ok this time around 2 minutes elapse on parity check and I'm past .4% says only 10 hours for rebuild. Guess I'll head to bed and check again in the morning. Hope it doesn't crap itself out and freeze up again. I tell you want, undraid is a adventure every time you try to change anything. tower-diagnostics-20230613-2309.zip

I reverted all my upgrades. Removed the nvme drive and the extra 32gb of ran I installed. I set the cache drive back to the 1TB SSD and formatted it. Hopefully it will rebuild now without crapping out again.

wow everything went to heck. I was restoring my backups then everything stopped responding. I checked the server itself and it said something about a critical Kernel error. rebooted stopped all dockers and vms system is running parity check now; estimating 9 days 47 errors found total raid size 4.11TB out of 21TB. not moving past .3% I think I'm done with unraid. every time I try to do something simple everything goes wrong. All drives reporting healthy and good temps. Browser interface stopped respoding. Here is picture of the server screen. Looks like the nvme is not responding. Please any thoughts. tower-diagnostics-20230613-2230.zip

Had to blow it all out and restore from backups. I just went with a single cache drive and used the 2TB nvme this time.

When I set the system to maintenance mode and do a Check Filesystem on the cache drive I get the warning device 2 is missing and errors found in extended allocation tree or chunk allocation. Looks like i have to blow out the whole thing and restore from a back up.

Old Cache drive SSD 1T I added a second drive to the cache pool nvme 2TB Raid1, And let it clone it self. Afterwards I tried to enable docker and VMs only to find neither of them would work. I tried pathing the vdisk to the cache drive I also tried pathing the libvirt.img location but nothing seems to be working. (Fix Common Problems) keep saying Unable to write to cache Any help would be greatly appreciated. I tried removing the second drive from the app pool, but then I got a error to many profiles. tower-diagnostics-20230613-2025.zip

Solved the issue. Had to turn on the following options in BIOS IOMMU - enabled SR-IOV - enabled In addition there are some values that have to be set to enabled putting them on AUTO seems to have no effect. Unfortunately I forgot what those were. But if you have the option of Enable use that rather than AUTO Everything looks stable and is running smoothly; there is still a peeked thread when someone joins but only for about 5 seconds then back down to 2-6%

I did notice that IOMMU is disabled on the board, maybe that will help some.

I had unraid running off a HP ML350 G5 2 processor at 2.3ghz and 12cores with 32gb of ram (slow ram like 660mhz) Either way I was having issues where the server and it would reboot randomly and I couldn't find the problem. I've been having issue with that server for years and I've put so much money into that I'm just fed up with it. Time for a simpler server and a quite one too. New system is running a Ryzen 7 2700k 8cores 3.7ghz with x470 Taichi and 32gb pc3200 The VM is running Ubuntu with 4 cores and 16gb ram We are currently running a 7Days server When players would join and play on the old server there was no lag but now on the new hardware we're experiencing quite a bit of lag. The VM cores are pinned to threads on the same core and what I'm noticing is that one thread will hit 100% then jump to another thread and hit 100% and it will keep hopping around keeping of of the 4 threads at 100% while someone is connecting and then it settles down. I didn't notice this behavior before on the old system. One thing to note; on the old system I was using onboard video and on this new server I have a really craptastic card I pulled from a hp minitower ati 200x or something. I just needed the card so the system would post. As far as I know the VMs video card is emulated so it should have nothing to do with my hardware one. Any thoughts?

JorgeB you nailed it. The cache drive was unmounted. And I am at fault for it. It was not caused by the upgrade but something I did. I had removed IP_Storage which consisted of 2 drives. But when I had removed this pool and must have removed that cache drive as well. I remounted the cache drive and my VMs and docker are back. My bad, Thank you very much JorgeB.

You know; i think your right; i think my cache drive is whats missing. I'll see if i can toss that back on or view its contents.

The 3 drives are from a old camera storage; I looked in them already.