Bridged NIC adapter always set to DOWN state
Description
Problem/Justification
Impact
SmartDraw Connector
Katalon Manual Tests (BETA)
Activity
Bug Clerk March 25, 2021 at 5:49 PM
Vladimir Vinogradenko March 25, 2021 at 5:49 PM
there are buggy cards/drivers where link state does not update immediately, that's why we are running dhclient regardless of link status
Chris Griffith March 24, 2021 at 10:22 PM
Good call that was it!
That might be worthy of it's own bug if it's trying to do DHCP on a port that doesn't have a cable connected to it?
Vladimir Vinogradenko March 24, 2021 at 10:11 PM
I think you still have DHCP configured on enp10s0f0 and it is waiting for a lease there and times out after a minute, hence the delay.
Chris Griffith March 24, 2021 at 9:40 PM
Attached, starting with "timeout-".
Just tried another reboot and did hit timeout again (image also attached) but was available sooner via ssh / WebUI (as soon as the "Console setup" screen was up).
I did simply cherry pick those two PRs code by hand into my system that is otherwise stock 21.02 alpha-1, so I am not at like "working branch" updates or anything if that makes a difference.
Details
Details
Assignee
Vladimir VinogradenkoReporter
Chris GriffithLabels
Impact
Components
Fix versions
Affects versions
Priority
LowMore fields
Time tracking
More fields
Time trackingKatalon Platform
Linked Test Cases, Katalon Defect Results, Katalon Studio Test Results
Katalon Platform
I am using a X540-T2 10GbE card with two interfaces, enp10s0f0 and enp10s0f1.
I used enp10s0f0 with DHCP (192.168.1.222) to setup TrueNAS. I then created a bridge, br0 with enp10s0f1 as the sole menber, and an IP alias of 192.168.1.88/24.
At this point I can connect to TrueNAS box directly WebUI / SSH via either IP address. So now I should be able to remove cable to enp10s0f0.
As soon as I do that, I can no longer access the box. `ip a s` shows that the interfaces are down.
enp10s0f1: <BROADCAST,MULTICAST> mtu 1500 qdisc mq state DOWN group default qlen 1000
br0: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue state DOWN group default qlen 1000
I bring up enp10s0f1 with `ip link set dev enp10s0f1 up` and do `systemctl restart networking` from the machine's shell via direct keyboard. A minute later, I can now access the box via the network again. Weirdly I can access it via both 192.168.1.222 (DHCP IP for the unplugged enp10s0f0) as well as the bridged static IP the cable is connected to.
After rebooting the box, the enp10s0f1 interface is in DOWN state again, and requires a manual upbringing and restart networking again to attach to the box (which seems to take another minute before I can access again). Which also means during restart the networking has to wait the full minute and a half to hit timeout before continuing for every restart.
I have also tried deleting the config for enp10s0f0 entirely with any effect.