[SOLVED] - Intel X520-DA not working properly with TrueNAS 13.0.1-U1 on MikroTik Switch

[lupuscon]

Hello everyone,

I am very new to TrueNAS, in fact this is my first setup.
The experience is not very good so far and i tried everything i could think of before writing this post.

Situation Summary:
System Summary
- HPE MicroServer Gen8
- Intel Celeron G1610T Dual Core
- 2x8GB HPE Smart Memory
- HDD Slots: 4x 960GB Kingston A400
- ODD Slot: 1x 240GB SanDisk SSD (TRUENAS CORE)
- USB Internal: 32GB USB Stick with a Grub Installation for Chainloading TrueNAS Bootloader
- PCIe Slot: Intel X520-DA with genuine Intel 10GBE-SR Tranceivers

• Fresh True NAS Core Installation on 240GB SSD with 16GBs Swap
• Single NIC Configured
• WebUI is not accessible
• TrueNAS not pingable even in the same VLAN
• lspci shows the NIC with the correct chipset and spec

What has been tried?:

• Trying different LAG Configurations
• Removing LAG and using Single NIC Port
• Configure via DHCP -> DHCP Client aquires an ip address and as soon as the ip is configured network connectivity is broken.
• Swapping NICs (I used a QLogic card before hand, found out TrueNAS does not support this controller) -> Switched to Intel X520s

Any Ideas are highly welcome.
Thank you in advance.

 

[Samuel Tai]

I wouldn't be surprised if your X520 is overheating or pulling too much power from the MicroServer's power supply. Is there any reason you're running 10G and LAGG at that?

 

[lupuscon]

Currently i am not even using LAGG as i stated in my situation report.
I tried to because i ultimately want to use an iscsi target on my truenas with my esxi server.

The card is not even getting handwarm.
Power usage seems normal. I already had a cpu with double the power consumption running without issues. (E3-1220v2) but i switched back to the Celeron G1610T to save some energy.
There is no Power Cap configured.

The behavior is very weird as soon as there is an IP configured the IP stack seizes to function properly even in the same subnet. I already checked.

 

[lupuscon]

Found the core issue
The issue laid within my switch configuration.
I did not suspect the switch, since DHCP was working and the VLAN seemed correct.

The core issue was "Independent VLAN Lookup" was not enabled in my MikroTik SwOS.

This is necessary if you have tagged uplinks/trunks and access ports on the same switch.

 

[macatarere]

Thanks for posting the cause and solution, I'm using MikroTik where I can, just starting on VLAN configuration, there is a lot to learn, an '8S+ is waiting to be patched in. Been very happy with Free/TrueNAS, switched from Nexenta, ~2010. This forum has been very helpful, good luck with your installation.

 

[Volts]

The core issue was "Independent VLAN Lookup" was not enabled in my MikroTik SwOS.

This is necessary if you have tagged uplinks/trunks and access ports on the same switch.

That implies that either the same MAC address is in use in two different contexts, or something is bridging traffic between interfaces/VLANs.

Have you been able to identify the underlying issue?

 

[lupuscon]

That implies that either the same MAC address is in use in two different contexts, or something is bridging traffic between interfaces/VLANs.

Have you been able to identify the underlying issue?

The issue might be the static Teaming between my FortiSwitch 148F and the MikroTik Switch.
However if i understood the flag correctly it was just because i mix access ports and trunk ports on the same switch.

P15/P16 beeing my static trunk Uplinks to my FortiSwitch P51/P52 with tagged VLANs only.
While having Untagged Ports P11/P12, as well as P13/P14, both pairs running LACP 802.3ad.

I think it is a MikroTik speciality to be honest, just like the PVID on Netgears ProSafe Series.