FumblingintheDark Posted October 20, 2019 Share Posted October 20, 2019 (edited) When I first built this server I was always under the impression the overclock was applied and working. However due to some issue with USB shorting out that caused the server to not boot at all, it was offline for a few months. I made due till recently and decided to get it back up and running. In the trouble shooting process I cleared the CMOS. I figured out that it was the USB when I removed the server from the rack and set it up on my desk. It booted first try. I used an extra SSD with a windows install to test all the hardware. I reapplied the OC and updated the BIOS. I continued to stress test it for about a week until I was confident it was stable. I have now reinstalled the server into the rack. Unfortunately when I try to start the array in unraid the server will crash and restart. Any overclock even the simple multi core enhancement will result in the array failing to start. It's late now but tomorrow I plan on trying to start the array with VMs not auto starting to see if it's a VM issue. I was also thinking about starting fresh as I know early unraid involved messing around with XML files for CPU pinning and GPU pass through as well as random other tweaks that I've made to get some random thing to work. All thoughts and guidance welcome with the exception of "overclocking = unstable = bad idea". If I can't get it to work in the end, I'll just have to suck it up, but I'm going to try! Let me know if there is anything I can provide to assist in troubleshooting. Thank you in advance! Hardware: CPU - 7980xe MOBO - Fatal1ty X299 Professional Gaming i9 GPUs - 1070, 1080ti Unraid ver 6.7.2 Edited October 20, 2019 by FumblingintheDark Quote Link to comment
John_M Posted October 20, 2019 Share Posted October 20, 2019 3 hours ago, FumblingintheDark said: "overclocking = unstable = bad idea" It's the truth and the sooner you admit it the sooner you'll fix your problem. Linux isn't Windows and server grade hardware is not overclocked for a very good reason. Quote Link to comment
Vr2Io Posted October 20, 2019 Share Posted October 20, 2019 (edited) Does once not OC then everything run normal ? ( include VM, GPU passthrough ... ) 4 hours ago, FumblingintheDark said: All thoughts and guidance welcome with the exception of "overclocking = unstable = bad idea". Really like this and support !! I seldom O.C. in recent years, mainly bacause O.C. always oppsite to power draw and not much gain. But some tweak I also will do. I belive you could identify the cause, because overclocker always harder for try and error. 🤣 Edited October 20, 2019 by Benson Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 5 hours ago, Benson said: Does once not OC then everything run normal ? ( include VM, GPU passthrough ... ) Currently with no OC everything works. I just wonder if in the past when trying to get turbo boost or some other out of the box thing to work I added a line of code or something that is causing the issue. Unraid has come so far that a lot of things you had to mod a file is now a simple mouse click. Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 7 hours ago, John_M said: server grade hardware I wouldn't call it sever grade. It's just a server in function. Quote Link to comment
John_M Posted October 20, 2019 Share Posted October 20, 2019 19 minutes ago, FumblingintheDark said: I wouldn't call it sever grade. It's just a server in function. I know. It's consumer grade. Even more reason to keep it in spec if you expect reliability. Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 (edited) I've added some files to hopefully help in the troubleshooting. I've noticed w/o the OC on the "watch -n 1 grep MHz /proc/cpuinfo" will show the speed get to 3.9GHz with the very rare 4GHz. I really have no idea why those are the chosen speeds seeing that the 7980xe has a base of 2.6GHz, Max Turbo of 4.2GHz, and a Turbo 3.0 of 4.4GHz. I've seen sites say the all core boost is 3.4GHz syslog.txt lscpu.txt syslog Edited October 20, 2019 by FumblingintheDark Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 tower-diagnostics-20191020-1615.zip Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 I noticed in System Profiler that it states max speed 4GHz. Is this coded into unriad and is it modifiable? Quote Link to comment
Vr2Io Posted October 20, 2019 Share Posted October 20, 2019 You complaint CPU clock speed always too high ? Not reach turbo freq ? Not keep in low when light loading ? or whatever .... what is the problem ? There are CPU Scaling Governor which you can control in OS, you may install "tips and tweaks" plugin to control that. But you also need to check does BIOS have any override or hard setting. Nowadays CPU are complicated, for example, some CPU have turbo in different form, i.e you load 2 core, then those core will be 4.9G, if loading 4 then 4 core 4.6G, if load all then core will be 4.3G etc. But some CPU also allow set all core in full speed all-time in BIOS. Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 The issue I'm having is they if I try to adjust clock speeds what so ever unraid will crash when I start the array. I can boot into unraid, I can even start the array in maintenance mode. I'm trying to figure out why this is. Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 So I was able to run the server OCed on a fresh install of unraid. I took 3 random drives I had laying around to create an array and was able to start and stop it with out it failing. I also verified that the CPU was reaching 4.6GHz (the OC) in unraid. Now I just need to figure out what configuration is causing it to crash. Any ideas on where to look? Quote Link to comment
Vr2Io Posted October 20, 2019 Share Posted October 20, 2019 (edited) So, you mean same OC setting haven't crash in simple Unraid config. Pls provide what OC setting in BIOS ? First think, if "BCLK" OC that may affect I/O clock and cause problem. Edited October 20, 2019 by Benson Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 So I don't think it's the OC setting. The reason is that my current unraid config crashes however, a fresh install doesn't. I've also tested and I'm confident that OC is stable after a lot of testing in bare metal Win10. I think it's a issue with a config in my primary unraid install but I'm unsure what it might be. I might have narrowed it down as when I copied over network configs, drive assignments, drive settings, and share settings the server started to crash again. Quote Link to comment
FumblingintheDark Posted October 20, 2019 Author Share Posted October 20, 2019 What files should I share? I can post a copy of my flash. What files should I remove before posting the flash drive backup? Quote Link to comment
Vr2Io Posted October 21, 2019 Share Posted October 21, 2019 (edited) 2 hours ago, FumblingintheDark said: The reason is that my current unraid config crashes however, a fresh install doesn't. No, OC also a factor because without OC, old config run normal, otherwise it should no crash whatever what config. I prefer know what OC have done, so may have some clues. Or you may try stop application one by one in old config, or add application one by one on new setup to found out the sweet spot. Edited October 21, 2019 by Benson Quote Link to comment
FumblingintheDark Posted October 21, 2019 Author Share Posted October 21, 2019 It's a 4.6GHz OC on all cores. I used this "guide" if you will.... It's almost embarrassing how easy it was and how stable it is when I tested it. Quote Link to comment
Vr2Io Posted October 21, 2019 Share Posted October 21, 2019 (edited) https://www.kitguru.net/components/leo-waldock/intel-core-i9-7980xe-extreme-edition-18-cores-of-overclocked-cpu-madness/2/ In page 2, OP also state most software fine except Adobe Premiere, would you try O.C. but stop at 4.2GHz ?? ( Power draw also reduce 100w ) At 4.6GHz all the cores appeared stable in the software we used with the notable exception of Adobe Premiere which caused the system to crash at 4.5GHz-4.6GHz and refused to open a Premiere Project at 4.3GHz-4.4GHz. Reducing the speed to 4.2GHz resulted in a system that ran Premiere perfectly. 36x100MHz, 1.000VID 37x100MHz, 1.030VID 38x100MHz, 1.050VID 39x100MHz, 1.077VID 40x100MHz, 1.100VID 41x100MHz, 1.124VID 42x100MHz, 1.148VID 43x100MHz, 1.175VID 44x100MHz, 1.203VID 45x100MHz, 1.203VID 46x100MHz, 1.203VID 4.7GHz system froze Default clocks, 1.005VID, 275W total 36x100MHz, 1.000VID, 310W total 37x100MHz, 1.030VID, 330W total 38x100MHz, 1.050VID, 350W total 39x100MHz, 1.077VID, 365W total 40x100MHz, 1.100VID, 390W total 41x100MHz, 1.124VID, 415W total 42x100MHz, 1.148VID, 450W total 43x100MHz, 1.175VID, 485W total 44x100MHz, 1.203VID, 525W total 45x100MHz, 1.203VID, 540W total 46x100MHz, 1.203VID, 550W total Edited October 21, 2019 by Benson Quote Link to comment
FumblingintheDark Posted October 21, 2019 Author Share Posted October 21, 2019 Interesting. I did not test Adobe Premiere. I'll test my current unraid set up with it to see if anything changes. If not I think I'll just have to start from scratch and rebuild. Quote Link to comment
FumblingintheDark Posted October 22, 2019 Author Share Posted October 22, 2019 I wasn't able to work on the server today. I'm curious if there are any settings that might be in the config that could cause an issue. 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.