Issue: 2000 Oversize packets per second on SG220-50 trunk port

[daltrucks]

GE49 on our SG220-50 is connected to EdgeRouter 8-Pro with multiple tagged vlans. This interface is recording a lot of oversize packets.

MTU on EdgeRouter is set to 1500 (for interface and VLAN vifs). I have tried enabling Jumbo Frames on SG220 and setting MTU to 9000 on EdgeRouter, but oversize packet count still increases at very high rate. With this increase the input errors counter goes up as well.

CISCOSWITCH#show interfaces GigabitEthernet 49
GigabitEthernet49 is up
 Hardware is Gigabit Ethernet
 Auto-duplex, Auto-speed, media type is Copper
 flow-control is off
 back-pressure is enabled
 238938 packets input, 293855180 bytes, 0 throttles
 Received 377 broadcasts (0 multicasts)
 0 runts, 185612 giants, 0 throttles
 185612 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
 0 multicast, 0 pause input
 0 input packets with dribble condition detected
 227313 packets output, 43180823 bytes, 0 underrun
 0 output errors, 0 collisions, 0 interface resets
 0 babbles, 0 late collision, 0 deferred
 0 PAUSE output

[Philip D'Ath]

How do you know they are oversize packets?  All that says is "input errors".

[daltrucks]

RMON oversize packets count matches up with input error packets pretty close and increases at similar rate. Approximetly 2 seconds between those two commands below were executed.

CISCOSWITCH#show rmon statistics interfaces GigabitEthernet 49
==== Port gi49 =============================
etherStatsDropEvents : 0
etherStatsOctets : 3222163328
etherStatsPkts : 20125809
etherStatsBroadcastPkts : 6073
etherStatsMulticastPkts : 0
etherStatsCRCAlignErrors : 0
etherStatsUnderSizePkts : 0
etherStatsOverSizePkts : 18030168
etherStatsFragments : 0
etherStatsJabbers : 0
etherStatsCollisions : 0
etherStatsPkts64Octets : 387895
etherStatsPkts65to127Octets : 220938
etherStatsPkts128to255Octets : 264021
etherStatsPkts256to511Octets : 106642
etherStatsPkts512to1023Octets : 117212
etherStatsPkts1024to1518Octets : 1554085

CISCOSWITCH#show interfaces GigabitEthernet 49
GigabitEthernet49 is up
 Hardware is Gigabit Ethernet
 Auto-duplex, Auto-speed, media type is Copper
 flow-control is off
 back-pressure is enabled
 20130000 packets input, 3228294325 bytes, 692 throttles
 Received 6073 broadcasts (0 multicasts)
 0 runts, 18034127 giants, 692 throttles
 18034819 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
 0 multicast, 0 pause input
 0 input packets with dribble condition detected
 10147503 packets output, 1746514431 bytes, 0 underrun
 0 output errors, 0 collisions, 0 interface resets
 0 babbles, 0 late collision, 0 deferred
 0 PAUSE output

[Stenli323]

Please, solved this problem? I have same problém. Is very critical and cisco support is passive.

[daltrucks]

Check bug report CSCuy16641

It is probably related to the issue. For us the issue has been more and more apparent, with noticeable VoIP issues. Only solution I see so far is changing to a different switch.

[Stenli323]

Ou no, the switches is half yare old, and we have bought 3... :-( I must test every night

[daltrucks]

There was another person having similar issue that were resolved by checking negotiation terms between between switch and router (duplex and speed). If settings match on both ends, even if in auto mode, then try configuring it manually on each end. I have yet to try this, but fail to see how it could be related to the issue.

This is really a joke. File transfer speed to/from devices on SG220 is 40MB/s tops, yet once you leave SG220 alone it is up to the max. Maybe I can play around with switch on weekends and see if resetting to factory defaults or changing some other settings will help.

[Tagesh]

hi, was this ever solved? im having similar issues.
deployed 7 SG220s 26p poe and all are having the same issue. Giants incrementing on all active interfaces. Additionally, after a few days, switch cpu utilization is maxed 100% and stays that way until you reboot.
Would really like some advice on this even if im a few years late.
PS. firmware upgrade is at the latest at the time of this comment.

[TorstenL]

Any news concerning this issue?

I have a new SG220-26P and RMON shows increasing oversize packets on all Dot1q Trunk-Ports and LAG. Enabling Jumbo-Frames doesn't make a difference. I have no input errors only the counter for oversize packets is counting up on the uplink and on the ports where access points are connected with vlans. On the other side of the uplink is a SG550XG with no errors and this SG550XG has multiple uplinks to other Cisco SG-Series switches (SG300, SG350X, SG500) where this issue doesn't appear, only on the SG220.

BTW: firmware version on the SG220 is the latest 1.2.0.6

Thanks.