Dear community,
We have a ISCSI network with MD3000i storage, R900 servers and switch DELL.PowerConnect 6248.
Some days ago we had a issue with a Virtual Machine and Vmware support informed us about high iscsi network latency.
Verifiying our virtual environment, the ping response from ESXs to MD3000i controllers have the following values:
(we are using Jumbo Frames)
ESX to MD3000i target
# vmkping -d -s 8972 192.168.130.160 -c 50
8980 bytes from 192.168.130.160: icmp_seq=24 ttl=64 time=1.475 ms
8980 bytes from 192.168.130.160: icmp_seq=25 ttl=64 time=1.416 ms
8980 bytes from 192.168.130.160: icmp_seq=26 ttl=64 time=1.442 ms
8980 bytes from 192.168.130.160: icmp_seq=27 ttl=64 time=1.503 ms
8980 bytes from 192.168.130.160: icmp_seq=28 ttl=64 time=8.358 ms
8980 bytes from 192.168.130.160: icmp_seq=29 ttl=64 time=1.469 ms
8980 bytes from 192.168.130.160: icmp_seq=5 ttl=64 time=15.902 ms
If in the ping test jumbo frames packets are ignored, the response is low but some times high (like jumbo frames test):
~ # vmkping -d 192.168.130.160 -c 100
64 bytes from 192.168.130.160: icmp_seq=27 ttl=64 time=0.300 ms
64 bytes from 192.168.130.160: icmp_seq=28 ttl=64 time=6.603 ms
64 bytes from 192.168.130.160: icmp_seq=29 ttl=64 time=0.402 ms
64 bytes from 192.168.130.160: icmp_seq=30 ttl=64 time=0.390 ms
Vmware support has recommended fix this issue, but all network devices are correctly configured.
My question is: Are normal those values high some times?
I think that those high values are normals because MD3000i has CPU process, but I dont know sure.
If anyone have any comments on this I would appreciate it.
Thanks!