SG350XG & SG350X Stack - ARP requests lost

[henrihoffmann]

Hi,

I recently replaced my core SG500X switch by a SG350X 48P & SG350XG 2F10 stack (recent firmware).

After a while I found a lot of issues, because of the unstable ARP operation of the switch. Sometimes the ARPs to remote devices (one more switch involved) worked, sometimes not. The core switch was allways able to ping the remote devices. But devices which are attached to the core had issues the ping via the core devices attached to a second stacked switch (2*SG550XG).

All entries (sh arp & sh mac address-table) where okay and all locally attached devices responded fine. To fix it, I powered off the SG350XG switch (unit 2), after that all issues are gone immediately.

Case 683419019.

Is there any better FW available?

Thanks 

Henri

[Ingrid Aragon]

Henri,

I have reviewed the case notes and I do not have a clear understanding what is the issue. I noticed that you have installed the latest firmware on the SG350XG 2.3.5.63.

Please provide me a better description of the issue in order to assist you. if this is a known issue we will tell you what information do we need in order to escalate case to L2.

Regards,

[Ingrid Aragon]

I have consulted with our Tech Leader and he will take ownership of your case and provide you the best solution for this case. He will be contacting you soon.

[henrihoffmann]

Hi,

after 1 week the issue is gone after downgrading the FW.

The problem occurs only at the current version.

Btw. there was NO SUPPORT, just some questions if the issue reoccurs or is gone.

Very disappointed.

Henri

[Mark.Zhadan]

Hello Henri,

I have trouble like yours in stack with lost of mac addresses.
It's only on second and third switches in stack, first working without troubles.
Latest firmware, and one before.
Problem still not solved by Cisco.

Can you please provide version number of firmware, that worked (downgraded)?

Thank you.
Mark.

[henrihoffmann]

Hi,

the issues are still existent. 

Some fixes are:

- Do not use a LACP uplinks across both stack units (I currently use Spanning Tree for redundancy).

- After every switch reboot (even at firmware bugs) unplug and replug all Etherchannel link cables (both at once). This gives ARP a new chance.

Welcome in the sixties ....

Best regards

Henri