Hi,
We are having major performance issues with two brand new SG500XG-8F8T switches.
They are operating as iSCSI only switches towards one Nimble CS300 array (SAN) and one Dell R430 server (Vmware 6).
When we connect everything we have link on all ports, traffic is flowing as it should, we can ping with MTU 8972 from vmware host to both controllers on the SAN.
The problem starts when we are doing some load testing from the host.
Running IOmeter from a VM on the host (located on the SAN storage) random ports starts flapping almost emmidiatly.
Here is an example:
04-Apr-2016 20:08:13 :%LINK-I-Up: te2/12
04-Apr-2016 20:07:59 :%NSFP-I-SFPGibicDetected: te2/12 SFP port is present, module type - 10G BASE-SR
04-Apr-2016 20:05:01 :%NSFP-I-SFPGibicRemoved: te2/12 SFP port is not present
04-Apr-2016 20:05:00 :%LINK-W-Down: te2/12
04-Apr-2016 20:05:00 :%LINK-I-Up: te2/12
04-Apr-2016 20:05:00 :%LINK-W-Down: te2/12
Here is the configuration from port 2/12:
interface tengigabitethernet2/12
switchport mode access
switchport access vlan 201
We have tried both with and without flowcontrol enabled, doesn't make any difference.
We have tried switching optics (SR/LR) cables (different brands of OM3 and SM) and we have tried connecting both controllers to one switch (different vlans) but the error still remains (on both switches).
We have tried multiple firmwares without any improvements.
So we feel that we have tried everything we can think of, and now rely on your expertise and experience to try and solve this problem.
Any ideas?
As a sidenote we have connected the host directly to the controller, we then get 1GBps and 35k IOPS (32 workers 32K seq read).
Kind Regards
Magnus
