Hi all,

 

For some time now, I've had an issue with my Unraid server restarting when the parity check starts. It usually happens within the first 1% or so.

 

What I've tried to do:

 

• Removed UPS, thinking it might be a power-related issue, but the problem persisted even after disabling the UPS (docker/VM enabled).

• Disabled docker, left VM enabled, and the parity check ran smoothly.

• Disabled VM, left docker enabled, and the parity check failed again.

• Disabled both docker and VM, and the parity check ran smoothly.

 

So, it seems like it might be Docker-related, but I'm not sure why.

 

I'm a bit out of options, so I thought maybe someone else might have a good idea.

 

Thanks!

london-diagnostics-20240828-0832.zip

Server rebooting on its own is almost always a hardware problem, but if it only does it with docker enabled, try to see if it's a specific container causing the problem.

Can this maybe be related?

 

Aug 28 09:23:34 london root: Error response from daemon: error while removing network: network br0 id a2ef645d11f0be732eac145067de6e1189b88ecc338abf00eb194ccd8209552b has active endpoints
Aug 28 09:23:34 london root: Error response from daemon: network with name br0 already exists
Aug 28 09:23:34 london root: Error response from daemon: endpoint with name Adguardhome already exists in network br0
Aug 28 09:34:31 london root: Fix Common Problems: Error: Macvlan and Bridging found

 

Macvlan is not a problem with 6.12.13, and it would make the server crash, not restart.

Just want to say what I currently have done. I have swapped the motherboard, CPU and mem. I have two Unraid servers with the same hardware. Disks are not replaced. After starting up, I still have the same behaviour. So not sure if it is hardware related. On the other server I did not have this issue.

 

How do I know if it is a crash or not? It's shutdown immediately, no reboot page. Can it be the startup is done because it did go down, because the behaviour or setting is to start after having power? This is BIOS setting iirc.

 

I still need to test different containers up/down.

2 hours ago, Joentje said:

It's shutdown immediately

Do you mean it powers off?

So when I start the parity check, it will freeze after a few seconds but you cannot refresh the page. After this is starting up again.

 

The syslog will show this:

Quote

emhttpd: unclean shutdown detected

 

Currently I have VM enabled and Docker engine enabled, but without running containers. Parity check still running fine till now, needs some hours to finish.

Edited August 29, 20241 yr by Joentje
extra update

That suggests the server is rebooting on its own, and that's almost always a hardware problem.

1 hour ago, Joentje said:

So when I start the parity check, it will freeze after a few seconds but you cannot refresh the page

 

This would make me suspect a power related issue as when a parity check is starting up is likely to be when there is maximum draw on the power supply.   It can also sometimes occur if you are using power splitters to the drives and you split a PSU cable too many ways so you get voltage drops.

After some testing over time, after the suggestion of @itimpi, I checked the power. Here are some details about my setup:

 

Intel Core i3-9100

Fujitsu D3644-B

2x Kingston KSM26ED8/16ME

6x Seagate 2.5" 5TB drives

Mini-box picoPSU 90

Leicke ULL Power Supply 120 W

 

I guess that combined with everything running, power requirements have increased over time. The parity check would spin up all drives, which would eventually "kill" the server because I didn't have enough power.

 

I have now replaced the power supply with the following:

 

Mini-Box picoPSU-160-XT

LEICKE ULL 156W

 

The server has been running smoothly for some days/weeks now, including during parity checks.

 

All, thanks for the help!