HP1820 - Received Packets with Error

[rds_correia]

Hello.

We have a couple of Procurve 1820-24G-PoE+ (J9983A) running firmware PT.01.10.

We have noticed in the Support File that we have 8 ports with RX packet errors.

All 8 ports configured with vlans taggng (aka cisco trunk mode) either to deliver all vlans to another switch or to deliver all vlans to a vmware ESXi server.

Besides that, we notice some slow response on the VM's. Pings and file transfer work fine with decent speed but the web services working on some VM's tend to be very (a lot!) slow when being access from a desktop on another switch. This didn't use to occur with the previous switches we had (old 10/100 Cisco switches). Anyone noticed anything similar? Because I found this blog post: http://www.revk.uk/2015/08/received-packets-with-error.html

Cheers

[kkpdk]

I have a mix of original 1800, 1810, 1810v2, and now a couple 1820. In total, this site has ~40 18x0 switches.

Yesterday I configured two 1820s (PT.01.11). They are edge switches, with their uplinks on port 1, and now vlan tags on that port. They were installed a few weeks before now, and for those weeks they ran unconfigured, with no errors logged on port 1.

"Received Packets with Error" will count whenever a 1518 byte frame (full MTU 1500 packet) goes over the link. With the tag added it becomes 1522 (which is as it should be), but apparently the switch categorizes this as an oversize frame. Enabling jumbo frames is not a workaround, it still logs errors on >1518 frames. 

And today, the users on that switch complained. I have disabled tagging on those switches again, and the issue is gone.

Into the bone pile they go.

[lch2003]

Filed a case on the same issue in Dec 2015. They said MTU 1522 bytes is not supported in this condition. It seems there won't be a fix in the future. :(