I have a Mellanox MCX354A-FCBT (flashed from HP OEM) that I'm running in SR-IOV mode with 8 VFs. The host is using the PF as its primary interface and, at present, I only have one guest running a VF.
The guest is experiencing tons of problem with the VF. Connections will run at full speed for a variable amount of time then hang. When they hang I see messages like the following via dmesg (on the guest). I don't see anything unusual on the host.
It's always queue 4. I'm always able to open new connections and resume so this glitch does not kill the entire VF, just one active connection at a time.
Host OS: CentOS 8
Guest OS: Fedora 32
Any ideas? Bad card?
The guest is experiencing tons of problem with the VF. Connections will run at full speed for a variable amount of time then hang. When they hang I see messages like the following via dmesg (on the guest). I don't see anything unusual on the host.
Code:
[53005.122622] mlx4_en: eth0: TX timeout on queue: 4, QP: 0xb94, CQ: 0xd6, Cons: 0xffffffff, Prod: 0x3d7
[53005.275385] mlx4_en: eth0: Steering Mode 2
[53441.341698] mlx4_en: eth0: TX timeout on queue: 4, QP: 0xb94, CQ: 0xd6, Cons: 0xffffffff, Prod: 0x3d8
[53441.495372] mlx4_en: eth0: Steering Mode 2
[53472.061348] mlx4_en: eth0: TX timeout on queue: 4, QP: 0xb94, CQ: 0xd6, Cons: 0xffffffff, Prod: 0x3d7
[53472.217599] mlx4_en: eth0: Steering Mode 2
Host OS: CentOS 8
Guest OS: Fedora 32
Any ideas? Bad card?