Input Discards

[mcdowellster]

Hello,

I'm seeing 0.033x per second input discards on all my nodes in the cluster with a random jump to 0.1 per second. Its constant over the last 400 days of monitoring in check_mk and I see the same thing on the hypervisors.

I've upp'd RX and TX to 4096 on all node.
Unifi AGG switches (10G) for SAN and LAN/Trunking.
Flow control enabled OR disabled results in no change.

Any thoughts?

 

[pep.]

Hello,

I'm seeing 0.033x per second input discards on all my nodes in the cluster with a random jump to 0.1 per second. Its constant over the last 400 days of monitoring in check_mk and I see the same thing on the hypervisors.

I've upp'd RX and TX to 4096 on all node.
Unifi AGG switches (10G) for SAN and LAN/Trunking.
Flow control enabled OR disabled results in no change.

Any thoughts?

Did you ever figure this out? I'm seeing the same thing, even though my nodes use separate networks for cluster communication. The main trunk for the VMs (and management VLAN) has these input discards visible in check_mk.