6.9.0 - 10gbit only working when 1gbit ethernet is disconnected - Solved

[leech]

Hi,

 

i have a Problem with Unraid and 10gbit Ethernet.

I installed 10gbit cards in both of my unraid servers.

The cards are working and show 10gbit transferspeed in the dashboard.

The old 1Gbit onboard ethernet port is still connected for failover reasons.

 

When i now transfer data over the 10gbit interface i only get 1gbit speed (~125mb/s).

This is as long as the 1gbit ethernet port is up. If i shut down the 1gbit port i can transfer with 10gbit speed on the 10gbit ethernet port (~500mb/s due to SATA 6gbit).

I can reproduce this on both systems. (if 1gbit port is down, 10gbit is working, if i take the 1gbit ethernet port up 10gbit only works at 1gbit speed)

For testing i connect by IP address so i can rule out that the wrong interface is used.

 

What if i want to do loadbalancing between the ethernet ports? can i only use the speed of the slowest ethernet port?

Is this wanted by design?

 

BR

 

Edited March 23, 2021 by leech

[JorgeB]

You should use different IP subnets for both NICs, than always use the 10GbE IP for server to server transfers.

[ken-ji]

Also, since you mentioned failover, the 10gbit NIC and the 1gbit onboard are probably bonded (as part of the default config), the default bond uses the least common speed of all active ports. You should break it up into their own ethX devices

Edited March 22, 2021 by ken-ji

[leech]

Hm 10gbit is working now but Docker isn't anymore.

 

the Containers i have are working but i can not add any new ones.

 

Here an example when i try to install apache:

root@localhost:# /usr/local/emhttp/plugins/dynamix.docker.manager/scripts/docker run -d --name='apache' --net='bridge' -e TZ="Europe/Berlin" -e HOST_OS="Unraid" -e 'PUID'='99' -e 'PGID'='100' -p '80:80/tcp' -p '443:443/tcp' -v '/mnt/user/appdata/apache':'/config':'rw' 'linuxserver/apache'
Unable to find image 'linuxserver/apache:latest' locally
docker: Error response from daemon: Get https://registry-1.docker.io/v2/: dial tcp 23.23.116.141:443: connect: network is unreachable.
See 'docker run --help'.

The command failed.

 

 

Any Ideas why this happens?

tower-syslog-20210323-1005.zip

Edited March 23, 2021 by leech

[leech]

Found it!

Somehow the default route went missing, added 0.0.0.0/0 + GW and now it's working again