Stat of IB device 'mlx5_0' failed on working link

Infrastructure & Networking Switches and Gateways InfiniBand/VPI Switch Systems
1

We have on machine out of 18 identical ones which refuses to do ibstat, because

ibpanic: [8821] main: stat of IB device ‘mlx5_0’ failed: No such file or directory

However, the link is up and running, and ibstatus is happy

	default gid:	 fe80:...
	base lid:	 0x17b
	sm lid:		 0x0fc
	state:		 4: ACTIVE
	phys state:	 5: LinkUp
	rate:		 100 Gb/sec (2X HDR)
	link_layer:	 InfiniBand

This situation is invariant under reboots (and cable change).

What might be the matter here?

Cheers,
Thomas

2

Wrong!

The IB interface is actually mlx5_1, and ibstat mlx5_1 shows basically the same as ibstatus.

However, the first interface works as the Ethernet interface of the box, and it has lost is gid

~# ibstatus
Infiniband device 'mlx5_0' port 1 status:
	default gid:	 unknown
	base lid:	 0x0
	sm lid:		 0x0
	state:		 4: ACTIVE
	phys state:	 5: LinkUp
	rate:		 10 Gb/sec (1X QDR)
	link_layer:	 Ethernet

This is of course what makes ibstat cause a panic.

Nevertheless: this Ethernet port seems to work well (it’s the primary network interface for this machine).
No idea what happened to its gid…

3

Hi Customer,

ibstat/ibstatus is mainly used to check the interface status in IB mode.

You can test again with the latest MLNX OFED driver.

Enjoy it.

4

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.