CSMACD frame error. who can find the root cause from this output ?

[Davor_7]

recently, i received many error report about the "frame error on device CSMACD-X - EIY-X"(X=number,Y=A/B/C/D)

because the output is very large, let me just paste one of them here. i can see that there is 5 frame errors. can you see any other abnormal from this output? and what stannds for the "frame error"? what can i do next?
thanks!!

Device Counters EIA0:
Value Counter
----- -------
1090040 Seconds since last zeroed
17349416967 Bytes received
25500463491 Bytes sent
61095816 Packets received
60377829 Packets sent
13711788400 Multicast bytes received
8802853319 Multicast bytes sent
15823418 Multicast packets received
8094695 Multicast packets sent
0 Unrecognized unicast destination packets
2186254 Unrecognized multicast destination packets
0 Unavailable station buffers
0 Unavailable user buffers
0 Alignment errors
5 Frame check errors
0 Frame size errors
0 Frame status errors
0 Frame length errors
0 Frame too long errors
0 Data overruns
0 Send data length errors
0 Receive data length errors
0 Transmit underrun errors
0 Transmit failures
0 Carrier check failures
0 Station failures
0 Initially deferred packets sent
0 Single collision packets sent
0 Multiple collision packets sent
0 Excessive collisions
0 Late collisions
0 Collision detect check failures

[Richard Brodie_1]

A frame check error means that one of the Ethernet packets had a bad checksum. It's usually a hardware or wiring problem.

However, everything else looks fine and 5 errors out of 60 million packets (one error every two days, on average) received isn't _that_ much to worry about.

You say "you received many error reports". Do you mean some of the other interfaces are worse, or are you seeing some other kind of errors? You can always attach the full statistics.

[Davor_7]

yes. some of other interfaces report same errors.(attached)
and also i donot understand the difference between CSMACD-1 and CSMACD-4?

btw, from some manual, it says "Ethernet allows a 1 in 10(8) bit error rate", and the output above is"227355824358 Bytes received with 5 Frame check errors"
so i think this rate is not acceptable.

[Ian Miller.]

CSMACD-1 and CSMACD-4 are different ethernet cards configured for DECnet/OSI.

Your LAN is obviously troubled.

[Art Wiens]

Considering the amount of data that has passed (17GB rcv, 25GB snt), 5 frame errors doesn't sound too bad considering it could have been a small "burst" caused by someone kicking the cable or dropping a floor tile too hard.

Do you have connectivity issues or are you worrying about 5 errors in a log that's 1,090,040 Seconds (~12 days) old?

Perspective,
Art

[Davor_7]

Miler,
what kind of trouble you mean? how to find it out and solve it?
can you explain more about CSMACS-1/2/4? does "1" means "EIA","2" means "EIB","4" means "EID"? or just a coincidence?


Art,
"12 days" is because we just had a server reboot :)
and after this reboot, i found such frame error appears. and then i wanna find out the root cause. that's the background :)
so, your suggestion is to continue to monitor?

[Volker Halle]

Davor,

you need to issue the following command to find out, which physical LAN device is associated with each CSMA-CD station (and look at their counters):

$ mc ncl sho csma-cd stat * communication port, all counters

If these errors showed up immediately after booting this server, could it be that they are related to speed and duplex autonegotiations between the server and the switch ?

Volker.

[Davor_7]

thanks Volker,
but our SAN is EMA
it displays an error when doing this command :)

[Volker Halle]

Davor,

what did you mean with this sentence ???


but our SAN is EMA
it displays an error when doing this command :)


We are talking about DECnet-Plus CSMA-CD stations or LAN network interface counters here (OpenVMS device name EIn0, DECnet-Plus CSMA-CD station csmacd-n). If I talk about a 'switch' in this context, it refers to the network switch, to which such LAN interfaces are typically connected and NOT to the Fibre Channel switch, which connects your OpenVMS host HBA to your EMA storage.

The command I've given is a NCL command to be issued on your OpenVMS node, which reported the CSMACD frame check errors.

Volker.

[Davor_7]

Hi Volker,
i tried your command, but following error displayed:
NCL>sho csma-cd stat csmacd-1 all chara
%NCL-E-CMLSENDFAILED, error sending command request
-CML-E-EMAAPROB, error returned from VMS EMA agent
-NCL-E-ACCESSDENIED, access denied
NCL>

[Allan Bowman]

As others have said, this looks like an awfully clean connection and 5 frame errors in 12 days is nothing to worry about. If it was truly a hardware or cabling problem, you would most likely see Carrier Check Failures in addition to the frame errors, and you would be seeing many more.

Allan in Atlanta

[Volker Halle]

Davor,


NCL-E-ACCESSDENIED, access denied


you need to issue these NCL commands from a privileged account.

Volker.

[Davor_7]

oh... thanks Volker, i missed this entry...
now i can see that "CSMACD-1" means "EIA"
BUT, can i know which PCI slot it is located at ?

[Volker Halle]

Davor,

$ ANAL/SYS ! needs CMKRNL priv & SET TERM/WID=132
SDA> CLUE CONFIG

will show the system configuration, it includes PCI slot information.

Volker.

[Davor_7]

oh! that's great!!!
thanks a lot, man:)

[Davor_7]

Hi Volker,
btw, i tried this:
SDA> clue fru
FRU Table written to CLUE$FRU.SYS file, use COMPAQ Analyze or DECevent to analyze.
SDA>

i got the SYS file, but i cannot get the correct output with ANALYZE command.
do you have any idea?

[Martin Vorlaender]

Davor,

COMPAQ Analyze and DECevent (and WSEA, part of WEBES) are layered products that must be installed separately.

cu,
Martin

[Volker Halle]

Davor,

you can download DECevent and/or WEBES (contains SEA = previously Compaq Analyze) from the Analysis Service Tools page:

http://h18023.www1.hp.com/support/svctools/

These tools need to be installed, but do not require a license for errlog translation.

With DECevent, you can then use:

$ DIAGNOSE CLUE$FRU.SYS

With WEBES, it's more convenient to use the WEB interface http://your-vms-server:7902

Volker.

[Davor_7]

yes. i used DIAG for the analysis :)
thanks

but it seems that the CLUE$FRU.sys output is not similar with >>>sho fru, hehe.

[Volker Halle]

Davor,

there may be another level of mapping done by the console level between the actual HW config and what OpenVMS is seeing.

You can also look at what the console reports during init (as extracted from DIAGNOSE CLUE$FRU.SYS).

With the output of CLUE CONFIG, >>> SHOW CONFIG and the view of the back of your system, you should be able to figure it out...

Volker.

[Volker Halle]

Davor,

sorry, my previous reply should have been directed to your SDA> CLUE CONFIG thread.

Volker.