| Title: | DEChub/HUBwatch/PROBEwatch CONFERENCE |
| Notice: | Firmware -2, Doc -3, Power -4, HW kits -5, firm load -6&7 |
| Moderator: | NETCAD::COLELLA DT |
| Created: | Wed Nov 13 1991 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 4455 |
| Total number of notes: | 16761 |
DEChub 900MS
DECrepeater 900TM
Alpha 3000 series systems connected to two ports
PC's connected to other ports (at least 6 PC's)
Counters for the ports have the item "Frame Check Sequence Errors"
Only on the ports connected to PC's do we ever see a positive value.
Number range from 100 - 200 with the Readable Frames value between 1
and 2 million. Basically a pretty small number. No operational and/or
performance problems.
Basic question from customer is what does this mean? and should I be
concerned??
From the Context-Sensitive Help:
"This counter is incremented by one for each frame received on this port
with the FCSError signal asserted and the FramingError and
CollisionEvent signal deasserted, and whose OctetCount is greater than
or equal to minFrameSize and less than or equal to maxFrameSize (refer
to Section 4.4.2.1, IEEE (see Section G0.70) 802.3 Std)."
Wheww..... My summary is that this is a frame whose FCS (Frame Check
Sequence) does not match the value calculated by the receiver, but the
frame is otherwise ok. Without having the actual spec in front of me,
I am not sure if the FCS field is also referred to as CRC-32. Some of
the diagrams I've been looking at imply that it is in the same location
within an Ethernet 802.3 CSMA/CD frame.
Am I looking at this correctly? Any paticular reason why these errors
are occuring? Should the customer be concerned?
All input greatly appreciated!
Thanks
Randall Buck
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 2719.1 | NETCAD::HERTZBERG | History: Love it or Leave it! | Thu Sep 07 1995 11:23 | 30 | |
>> ...this is a frame whose FCS (Frame Check Sequence) does not match
>> the value calculated by the receiver...
Correct.
>> ...but the frame is otherwise ok.
The FCS error just means there is an error somewhere in the frame. It
_could_ be in the FCS itself, but this is a pretty small probability,
considering the size of the FCS field compared to the size of the rest
of the frame.
>> Any paticular reason why these errors are occuring?
Hard to say from a distance. Definitely not a generic design flaw in
the 900TM. Possibly a faulty 900TM. Switch the Alpha and PC ports to
see if the error stays with the stations or with the ports. Most
likely cause is the NIC cards in the PCs are a little shaky.
Hopefully, they're not ours... ;^)
>> Should the customer be concerned?
I would be, only because this is not a normal situation. No danger of
corrupted network data... there are plenty of layers of protocol
guarding against that, but every FCS means a retransmission of the
faulty packet. This has the effect of degrading network performance,
and the effect is worse than what you might guess by looking at the
percentage of corrupted packets. Are the PC users complaining about
their network performance?
| |||||
| 2719.2 | Only the PC's | CSC32::R_BUCK | Have been assimilated | Thu Sep 07 1995 14:28 | 15 |
Thanks for the update and confirmation. More/updated information:
Customer is pretty sure the error follows connection to a PC. Will try
to test this later to be absolutely sure.
All PC's are similar hardware. All use an Etherlink 3, (Etherlink
III?), adapters from 3Com. Not sure of the specific model or models.
Pathworks V5.x is the primary application using the network.
Quick check has not turned up any specific problems with 3Com adapters
and Pathworks. Customer did not have any specific complains about
performance although he does believe things slow down during periods of
high usage.
Randall Buck
| |||||