SOLVED FreeNAS / hardware is shutting down

[skicek]

Hello everyone,

*********
Solved on 99%. Issue is with C6 state, on ryzen first and second generation. Found 3 ways to solve issue:

1. Write script and add it to cron to be run every 1 minute
2. change in bios Power state idle to typical.
3. disable C6 in bios

It would be good if someone would add this info in some support documentation or something, as looking on google results lots of people have this problem.
*********

Im new to FreeNAS software and I have a problem with not quite sure what exactly... I will try to describe as best as I can.

Hardware:
AMD Ryzen 3 1200, 3.1GHz, 8 MB, BOX (YD1200BBAEBOX)
Cooler Master MasterWatt Lite 500W (MPX-5001-ACABW-EU)
ASRock AB350 Pro4 (NIC Realtek RTL8111GR )
UPS Green Cell UPS01LCD
SSD Intenso TOP 256GB SATA3 (3832440)
Kingston DDR4, 16GB, 2400MHz, ECC (KTH-PL424E/16G)
Western Digital Red 4 TB 3.5" SATA III (WD40EFRX)

Configured with your guides (dont remember who wrote it)

TL;TR:
Everything was working for 12 days, now system is going down every 1 day.
No response from network (not visible from router)
Checked logs no entries there (one hour before system down there is only entry "(DEBUG) EtcService.generate():274 - No new changes for /etc/krb5.conf")
HDD temp is 40*C (alert set for 50, but never receive one so guessing it is not going that hot)
CPU usage, system load diagram is missing from time when it going down from network.



So when I assembled it and installed FreeNAS on SSD (I know to much space wasted for OS...) everything was working fine by 12 days after that problem start appearing... problem is that machine is working (powered) but system is loosing connection to network (Static configuration). That was I was thinking on first as I couldn't connect to it... but then I start debugging, checking logs (there is nothing on logs - /var/log - no entries on the time when machine is disappearing) but then I was wait maybe temperature is to high so I went to reporting and see that everything is fine... but is it? no I saw that when machine disappear there is no CPU Usage, no CPU temperature recorded, no system load... and this is the point where I dont know what to do more or how can I find root cause of it. Machine is in headless mode, and now it is happening almost every day.

So I hope someone had similar problem or have any ideas what to do next. Hardware is connected to UPS which is not detectable by FreeNAS unfortunately :( but also my router is, so I can remove UPS as a problem because if it would also router should shutdown but that doesnt happens.

 

[Commissar]

it's the realtek nic. i had the same issue with the b450 pro4 nic of a similar model.

 

[skicek]

but how realtek NIC can affect whole system like that. I could understand that if it was only the case that system is not visible on network but, no like there is no use of CPU etc...

For me it looks like system is down but rest of the component like fans are still powerup. I have somehow similar issue on my PC where im gettin random BSOD and same no entries on sys logs ete only thing I know is that system ssd is not powerup and I need to discharge all power coils to start PC again... have this issue like from 5 years i replaced MB, sata cable and nothing helped on that but issue is so random that its hard to investigate as it can be once per year or sometimes few times per month.

On the NAS hardware the difference is that I dont need to discharge coils when rebooting, rest looks similar. So I highly doubt that this is realtek NIC, but I asked friend to check if he have some intel NIC so I could test.

 

[ThreeDee]

Well, if NIC is flaky, you will lose connection to it . .if you stick a monitor and keyboard on your FreeNAS box you could do more troubleshooting ..using the Intel NIC will probably cure your woes ..

Your PC though could have a bad PSU if you haven't tried swapping that out already .. a bad capacitor on anything, if it doesn't cause complete failure, can cause all kinds of random odd issues. In fact ..if you have flaky power in your house, that can cause all kinds of issues if your machines aren't on UPS's. (brown outs can be just as bad as a power surge)

 

[skicek]

About NIC, can you tell me if it will be ok if chipset will be Intel i350 or it has to be a manufacture by Intel with that chipset?

 

[JaimieV]

The chipset is the important thing, it doesn't need to be sold by Intel.
Do your future self a favour and take a look at the FreeNAS NIC hardware list and get something from that.

 

[skicek]

Yea I saw that list, but the problem is in my country it is hard to get recommended hardware, thats why I had to improvise. Ok so I getting that NIC D3035-A11 on chipset INTEL NHI350AM2 not 100% sure if it is the same as i350-am2?

as for now I changed some save power setting in BIOS and it is up for over 2days so Im still monitoring it

 

[skicek]

Finally I got NIC Intel i350, installed it and configured but problem is still there. After installing it was working around 10h after reboot it was working 4h...

DOES any one have any idea how to resolve this issue?

 

[Commissar]

Finally I got NIC Intel i350, installed it and configured but problem is still there. After installing it was working around 10h after reboot it was working 4h...

DOES any one have any idea how to resolve this issue?

do you have your system in debug mode?

but might be the PSU failing

 

[skicek]

yes, I have in system -general syslog set to debug.

PSU as everyother part is brand new, except intel NIC as I couldnt get new one so it is used one, but as I can see NIC is not a problem here as same problem was on Realtek and same on Intel

 

[skicek]

Im not sure how to read that... the missing parts in timeline is the time when I cant find device in network and reboot of it is needed. But the device itself is power up

 

[skicek]

Solved on 99%. Issue is with C6 state, on ryzen first and second generation. Found 3 ways to solve issue:

1. Write script and add it to cron to be run every 1 minute
2. change in bios Power state idle to typical.
3. disable C6 in bios

It would be good if someone would add this info in some support documentation or something, as looking on google results lots of people have this problem.

 

[05nelsonm]

Solved on 99%. Issue is with C6 state, on ryzen first and second generation. Found 3 ways to solve issue:

1. Write script and add it to cron to be run every 1 minute
2. change in bios Power state idle to typical.
3. disable C6 in bios

It would be good if someone would add this info in some support documentation or something, as looking on google results lots of people have this problem.

Was having the same issue on mine about every day or 2, requiring a restart. Ryzen 9 3900X [3rd gen]. After implementing suggestions 2 and 3, have not had a problem for a week now. Thanks!

 

[ChrisRJ]

HDD temp is 40*C

I would recommend to stay below 35 C. Statistics indicate that 40 C increases risk for failure after 3-5 years quite a bit.

 

[pschatz100]

Numerous posters have reported problems with the deep sleep states as implemented on the newer, gaming oriented, motherboards. The solution is to disable the power management in bios - or at least set a power profile that does not include deep sleept (C6.) If possible, I would also make certain the NIC remains powered up when the machine goes to sleep. Some motherboards offer an option for this in the bios.