cam94z28 - My Unraid server has random freezes

[cam94z28]

FWIW, I'm having a similar issue with a Ryzen 7-1700 on a MSI B350 Tomahawk Motherboard. It will run fine for many hours, then just randomly freeze. I had to run a parity check on one drive yesterday after bumping the USB stick (I know should move it to one of the rear ports), and temporarily wiping out the license info. I left it on last night, and today for the entire day at work. I came back to see the monitor still powered on (video signal), but the server totally unresponsive. Apparently it never finished the parity check in that time because it restarted after i restarted it. .

 

Many times it does not come back from a reboot, either. I have to shutdown, and manually power back on instead.

[trurl]

10 minutes ago, cam94z28 said:

I'm having a similar issue

Often the beginning of a discussion that turns out to be very different in the details. I have split your post into its own thread so the other user can have his thread and you can have yours.

 

Attach diagnostics to your NEXT post in this thread.

 

11 minutes ago, cam94z28 said:

Ryzen

 

[cam94z28]

Point taken. Thank you for the common sense support. Attached is my diagnostics file. Something else to add. I have my hard drives set to 15 minute spin-down delay. When the freeze has occurs, it's usually down to the parity drive and one remaining drive still online. The identical setup (other than RAM, Motherboard, CPU), was working perfectly on an FX-6300 with MSI 970A Krait SLI Motherboard.

antec300-diagnostics-20220415-2324.zip

Edited April 16, 2022 by cam94z28

[trurl]

Even better would be diagnostics taken with the array started.

 

Did you check out that link I gave about Ryzen?

 

If you are having crashes setup syslog server

[cam94z28]

I can't seem to find the original thread with the Ryzen link. I must have closed the tab. Unless you are referring to the syslog server link. I have just enabled syslog mirroring to flash. Hopefully I can catch the next crash.

 

Was the array offline? I will attach another diagnostic. Parity check is currently running. Hopefully that counts.

antec300-diagnostics-20220416-0028.zip

Edited April 16, 2022 by cam94z28

[trurl]

7 hours ago, cam94z28 said:

can't seem to find the original thread with the Ryzen link

In my first reply in this thread, there is a link to a post in the FAQ. Click on the arrow

[trurl]

Unrelated, but appdata, domains, system shares always have open files so your disks can't spin down with them on the array. And your docker/VM performance will be impacted with these on the array. Most people put these shares on cache or other fast pool.

[cam94z28]

15 hours ago, trurl said:

In my first reply in this thread, there is a link to a post in the FAQ. Click on the arrow

Thanks, I see it now. I must have missed the text that clearly said Ryzen.

15 hours ago, trurl said:

Unrelated, but appdata, domains, system shares always have open files so your disks can't spin down with them on the array. And your docker/VM performance will be impacted with these on the array. Most people put these shares on cache or other fast pool.

I have plenty of old SSD's laying around, so I definitely plan to install a cache drive (or pool). I have been avoiding it as there's nowhere to really mount them easily in this ancient case.

My server was basically idle, other than the parity check running. No docker containers or VM's running, and no active file transfers, so the drives did manage to power down. I came home today to an unresponsive server again. Cannot load the web interface, and it isn't responding to pings over the network. It still hasn't finished a successful parity check, due to freezing. Only option is to long-press the power button to power cycle.

 

I will attach the syslog I was mirroring to flash. I already looked it over, and there doesn't appear to be much that happened since I logged out last night. Maybe you will catch something I missed. Meanwhile, I will take a look at the BIOS for settings related to C-States.

EDIT: My BIOS Has "Global C-State Control". It also has "Power Supply Idle Control" with Auto, Low Current Idle, and Typical Current Idle as options. It was set to 'Auto'.  I'm assuming the latter is going to be the one to focus on?

 

Changed the "Power Supply Idle Control" as outlined in the FAQ. Crossing my fingers for a successful parity check.

syslog

Edited April 17, 2022 by cam94z28

[cam94z28]

Problem appears to be solved. Came home to a completed parity-check notification and a running server.

Thanks for the advice.