Loose 10GbE port causes ATA error count

[Chin-Fah HEOH]

I have recently installed a 300TB FreeNAS (11.2_U5). It is running on a 16-bay Supermicro and expanded to a 24-bay Supermicro JBOD via AOC-S3008L-L8E (which is basically an LSI 3008 HBA). The HDDs are Western Digital HGST 10TB (P/N: HUH721010ALE600)

There are 2 x 10GbE RJ-45 ports on the motherboard (Supermicro X11SPi-TF Mainboard), seen as ixl0 and ixl1 on FreeNAS. The ixl0 port is link aggregated (LACP) to another 10GbE port on a separate NIC card (AOC-STG-I2T, which uses Intel X540 chipset). So lagg0 members are ixl0 (on mainboard) and ix1 (on the NIC). The RJ-45 connector of the Cat 6 UTP cable to ixl0 is loose, and we get an intermittent flap of network disconnection ... nothing serious.

The funny thing is, on 2 occasions, when the network disconnection happened, the disk da25 (which is in the 24-bay JBOD) was affected (see attached file). The ATA count error of da25 increased from 0 to 1.

I am trying to wrap my head around this because I am puzzled why a loose 10GbE port would cause a particular disk to have an ATA count error.

Hope to get some advice/enlightenment of this strange phenomenon. Thank you in advance.

 

[dlavigne]

Were you able to figure out the cause for this?

 

[Chin-Fah HEOH]

Hi Dru,

I have no idea. Every time I jiggle the 10GbE port (it's still loose), the da25 (always that disk) ATA error count message came up. It is always 0 to 1, never more than that. I am avoiding touching the network cable from now.

If I find anything, I will update this thread. Just weird for me.

Thanks