cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
244
Views
0
Helpful
4
Replies

details of CSCuw44957

Hello,

Can someone please share the details of the bug CSCuw44957 ?

The workaround says  (..) It won't heppen on 15.0(2)SE8 or earier. (..)

But.

I have WS-3560X @15.0(2)SE4, and its ports work fine when in 1G mode.

When limited to 10/100 (auto), I found one tcp frame that gets dropped.

Nothing special, just something in the middle of TCP segment, but, the result is - the upper layers get confused and application hangs.

Maybe I'm hitting the same bug, and should consider migration to 15.2 train ?

Thank you in advance.

4 Replies 4

Ganesh Hariharan
VIP Alumni
VIP Alumni

Hello ,

Here are the details of the Bug for your information.

Symptom:
Some fragmented packets drop on the Cat3560X.

Conditions:
The issue is confirmed on 15.2(1)E,15.2(2)E2 or 15.2(3)E2.

It is not seen on 15.0(2)SE5,15.0(2)SE6 or 15.0(2)SE8.

Workaround:
It won't heppen on 15.0(2)SE8 or earier

============

Known Affected Releases:
15.2(3)E
Known Fixed Releases:
15.2(4.1.5a)E
15.2(5.0.43)E
Hope it Helps..
-GI
Rate if it Helps..

Mark Malone
VIP Alumni
VIP Alumni

Hey i cant obviously confirm whether your hitting that bug from what you provided but i can tell you the image your on is ED and the SE8 is an MD and the safe harbour image so Cisco recommend that image for that switch currently there most stable

I wouldnt bother going up to 15.2 either as there all ED as well.Always good to stay on a MD release.

Thank you for immediate answer, Mark.

I've switched to 15.0(2)SE8, the problem is still there.

During some innocent TCP transmission, one of the frames is lost, but only when port is NOT in 1G mode.

We also managed to rule out NIC drivers/model/manufacturer.

One thing thats changed: now I can see output drops on interface. Of course there is no telling if these are exactly these missing frames, but chances are high.

I guess now it's time to do some additional sniffing, and then move to 15.2.

Regarding my initial question - since the bug description is somehow 'generic', are there any further details ? what types of frames were lost, under what circumstances, how was this tested ?

Thank you

Hi no unfortunatly no further details Ganesh has actually posted the full output below from the bug page its very limited, to get your answer you could speak to TAC they probabaly wont have the answer either on hand but they can speak to the developers who should have more information regarding it and what exactly should be effected.The fact the MD release is also effected is not good hopefully they fixed it in one of the newer EDs

Review Cisco Networking products for a $25 gift card