Hello Networking,
I have a problem that's giving me a huge headache. I am running a highly sensitive test on some Macbook clients that pings a couple of IP addresses on the public internet over 24 hour period and fails if over 2% packet loss occurs. Now, the network path for these tests is as follows
Macbook --> Aruba MAS Switch --> Aruba controller --> ISP Modem(Internet_destination)
Here's the thing, when the test runs (and I have confirmed it is simply sending ICMP pings size 1024 over an interval of 1 second) the final result shows an average of 2-8% packet loss over the entire course of the test. To make matters worse, if I try to prove out the network by running the following ping tests I show <1% packet loss every time over extended ping tests (I have monitored 4-8 hours pings and shown 0% packet loss)
Ping SRC: Aruba Controller
Ping DST: Macbook
Path: Controller --> MAS Switch --> Macbook
and
Ping SRC: Aruba Controller
Ping DST: Internet_Destination
Path: Controller --> ISP Modem(Internet_Destination)
and finally, I have learned that if I hook the Macbook directly to the Aruba controller so that path is now
Macbook --> Controller --> ISP Modem(Internet_Destination)
I experience 0% packet loss on the test.
Any idea what I could be missing, or why I would be experiencing low packet loss when the ping's source from the Macbook instead of the Controller? I have captured Wireshark logs on the Macbook and validated that the pings are being sent out the interface. I do not currently have a way to validate that the pings are actually hitting the switch (MAS Switch doesn't have SPAN capability with our VLAN design.)
No comments:
Post a Comment