Unraid Randomly Dropping Internet

[Badboy]

Hi,

Unraid has been dropping the internet  Nothing else is dropping it.  I have Fiber line. It went a week without dropping it. but In the last couple days it has dropped it twice. Just a couple hours ago was the last one.

Here's what I have tried.

I ran it without the UPS, for a week.

Change the 10g Nic

Updated Bios

Reset Network Settings

Ran a ram test. Passed everything.

Tried Disabling Bonding.

 

Have not installed it yet, bought a new power supply, and a (CPU cooling system. Waiting on delivery.} Attached a couple files. Any ideas are appreciated. Ran this system for a long time with no major issues. I think it might be power supply.

Current power supply is a corsair 750W

Have 4 10TB hard drives, 16TB parity, 

1 GPU

1 NIC
 

Specs

Unraid server Pro, version 6.11.5

Model:Custom

Motherboard:ASUSTeK COMPUTER INC. - ROG STRIX X470-F GAMING

Processor:AMD Ryzen 9 3900X 12-Core @ 3.8 GHz

HVM:Enabled

IOMMU:Enabled

Cache:L1 - Cache = 768 kB (max. capacity 768 kB)

L2 - Cache = 6 MB (max. capacity 6 MB)

L3 - Cache = 64 MB (max. capacity 64 MB)

Memory:32 GB (max. installable capacity 128 GB)

DIMM_B1 = 16 GB,

DIMM_B2 = 16 GB,

Network:bond0: fault-tolerance (active-backup), mtu 1500

eth0: not connected

eth2: 10000Mb/s, full duplex, mtu 1500

Kernel:Linux 5.19.17-Unraid x86_64

 

badboy-diagnostics-20221207-1554.zip badboy-syslog-20221207-2051.zip

[JorgeB]

Nothing obvious to me in the log, do you know the time when it last happened?

[Badboy]

Hi,

Sorry for the late reply. I normally see an email when someone responds  been a little frustrated lately. 😁 It happened just shortly before I sent the post in. I have attached my network settings. I have this nic card https://www.amazon.ca/Asus-XG-C100C-Network-Adapter-Single/dp/B072N84DG6/ref=sr_1_1?crid=3R3T27OBBZ8EH&keywords=Asus%2BXG-C100C%2B10G%2BNetwork%2BAdapter%2BPCI-E%2Bx4%2BCard%2Bwith%2BSingle%2BRJ-45%2BPort&qid=1670697910&sprefix=asus%2Bxg-c100c%2B10g%2Bnetwork%2Badapter%2Bpci-e%2Bx4%2Bcard%2Bwith%2Bsingle%2Brj-45%2Bport%2Caps%2C75&sr=8-1&th=1  I have switched it with the identical one I have. Don't have a high skill set when it comes to this, but I didn't see anything that looked bad in the logs.  Could it be possible the USB flash drive is causing a disconnect? Not that old, think it is a Sandisk. Also, would the updates from Unraid possibly effect the nic card?

 

I replaced the power supply on Friday.  That night it disconnected twice from the network. I never make any drastic changes, thinking it might be the motherboard. I just bought that one last year.  Going to add a liquid cooling system to the cpu, and upgrade all the fans. Don't think that is the issue, but it won't hurt.  I have a Thermaltake W100 case.  Seeing the CPU and GPU run a high for what they are doing at times. Nothing too bad.  No fan in the front of the case to suck air in, going to change that.  Really don't think that is the issue, ran it this way for 5 years.  Been shopping for a board but not aggressively yet. Looks like a system rebuild shortly if I keep going. LOL...Thanks for getting back to me. Really appreciate it.

Edited December 10, 2022 by Badboy

[JorgeB]

Nothing obvious in the log that I can see, you are using eth2 as you main NIC (in a bond), technically it's OK but we recommend using eth0 as the main one, also since there' no link in the other NIC remove the bond, if it still happens suggest trying using the other NIC instead.

[Badboy]

5 hours ago, JorgeB said:

Nothing obvious in the log that I can see, you are using eth2 as you main NIC (in a bond), technically it's OK but we recommend using eth0 as the main one, also since there' no link in the other NIC remove the bond, if it still happens suggest trying using the other NIC instead.

The system totally crashed in the night, no power when I woke up. Ordered a new board. Should not be losing total power. That's twice now in the last number of weeks. Really strange. I thought I had put the eth0 back, guess not. Thanks for the help. Let you know what happens. Unless the UPS is part of the problem and I just missed it. I ran it for a week without UPS and it still dropped the internet, didn't do the crash part. 

[Squid]

While you have set the memory speed correctly, have you also set in the BIOS Typical Idle Power (or similar wording) to be Typical (or similar)

[Badboy]

1 hour ago, Squid said:

While you have set the memory speed correctly, have you also set in the BIOS Typical Idle Power (or similar wording) to be Typical (or similar)

I  have never looked at that setting. Have had this board over a year with no issues up until the last number of weeks. I can have a look and see what it is set at.  I think most of those settings were auto. Could be wrong. I will have a look and let you know. I hope I never see a power or network connection issue again. It's frustrating, not to mention expensive. Positive side system should run more efficient. Not what I really wanted, but ended up buying the Gigabyte X570s Master. Not a lot out there right now. For me to go to an actual server board would probably mean changing the cpu and ram. Just could kick myself for not just changing the board from the start. Thanks, let you know. 

[Badboy]

Just a follow-up, ended up changing the board out, solved the power issue.  That was pretty much a rebuild LOL.  The new problem is Unraid is dropping the internet every other day. Squid mentioned the idle settings in his previous post. He might be right about those settings. I haven't had a chance to look yet. I know I have to go and change the keyboard and mouse settings, can't boot without them plugged in. Some good come out out this, the server runs much quieter with the new fans and the Artic cpu cooler, which brought the temperature down pretty good.