SUMMARY: ie0: no carrier

From: M Applebaum (applebau@sunking.aoe.vt.edu)
Date: Fri Sep 18 1992 - 02:48:53 CDT


  Sorry it took so long to get this summary back to all of you.
  Again, thanks for your help.

  This summary is kind of long. To shorten it a bit I removed the
  portions where people recopied my original question (except where
  it would make the response to the question confusing). The original
  question has been placed in the summary to make it complete.

  Thanks for the help

  Mike Applebaum
  applebau@sunking.aoe.vt.edu

===============================================================================
                           ORIGINAL QUESTION
===============================================================================

  We have a Sun 3/280 running SunOS 4.0.3.

  The problem: we are constantly getting
    "ie0: no carrier"
  errors from this machine. However, no other machine
  on the same ethernet segment is having any problems.
  This is the first machine on the segment.

  Sun came out and replaced the CPU board (with the
  ethernet controller). I've replaced the transceiver
  and transceiver cable twice (with ones that worked
  fine on the other machines on this segment).

  Our engineer from sun suggested it may be a software
  problem or a problem in the network. Since this is the
  first machine on the segment and no other machines on this
  segment are experiencing problems, I doubt its in the
  ethernet itself (but I could be wrong). Does anyone have
  any clue as to what may be causing this problem.

  This problem is extremely important to us because this particular
  machine is our primary NIS/communications/disk server.

  Thanks,

  Mike Applebaum
  applebau@sunking.aoe.vt.edu
           [128.173.7.101]

===============================================================================
                              MESSAGE 01
===============================================================================

>From davies@gest20.SINet.SLB.COM Wed Sep 9 03:00:39 1992

I don't know your set up mike, but if you have a backbone with a transciever via multidrop to your machine.

ie ------------------------------------[]----------------
                                         +
                                          +
                                         MACHINE

make sure that the termination is correct on the backbone - bad termination causes reflected signals; these will decay in amplitude away from the termincation point. If you have a bad connector/terminator at the end you may get enough reflection to corrup
t the signal at the point where it is tapped near the end but not really affecting it further away. check that bit out!

Gareth
 
===============================================================================
                              MESSAGE 02
===============================================================================

>From fetrow@biostat.washington.edu Wed Sep 9 03:17:44 1992

 Well, the first thing to try might be physically shoving a different
machine in there and see if it has the same problem (using the same
cables). PC's are really convenient for this.

-- 
 - dave fetrow-					fetrow@biostat.washington.edu

"Never underestimate the ability of large potential profits to make technically inferior technology into a workable standard."

=============================================================================== MESSAGE 03 ===============================================================================

>From alastair@eucad.co.uk Wed Sep 9 06:18:50 1992

Try replacing the terminator, or adding an extra length of cable between the transceiver and the terminator.

Just a thought

Al

--------------------------------------------------------------------------- Alastair Young Systems Supervisor (SMTS) _ _ Ariel NH European CAD Developments Limited aka EuCAD )/___ (_) Red Hunter EuCAD/Cadence mail: alastair +44-31-225-3434 __/(___)_*##/c "Aspencade" Full Email Address: alastair@eucad.co.uk / /\\|| \ / \ If in doubt 9 Wemyss Place, Edinburgh, Scotland EH3 6DH \__/ ----'\__/~~~ Paint it RED ---------------------------------------------------------------------------

=============================================================================== MESSAGE 04 ===============================================================================

>From pjw@math30.sma.usna.navy.MIL Wed Sep 9 07:25:19 1992

We had a similar problem with a 4/370. The 2nd ethernet interface looks like its powering the xcvr etc but is not making a good connection. We replaced the interface, etc. It turned out the solution was removing a washer from the AUI connector, to get a tighter fit. ---------------------------------------------------------------------- Dr. Peter J. Welcher EMAIL: pjw@math30.sma.usna.navy.MIL Mathematics Department, M/S 9E PHONE: (410) 267-3606 572 Holloway Road U.S. Naval Academy Annapolis, MD 21402-5002 FAX: (410) 267-4883 ----------------------------------------------------------------------

=============================================================================== MESSAGE 05 ===============================================================================

>From heiser@tdw901.ed.ray.com Wed Sep 9 09:09:01 1992

I know this doesn't help much ... but every time I've seen that "no carrier" problem, it has been due to a problem on the network. Check for things like missing/loose terminators, loose connections, etc.

-- Bill Heiser Work: heiser@sud509.ed.ray.com Home: bill@unixland.natick.ma.us Public Access E-Mail, USENET, UNIX, and UUCP Accounts 508-655-3848

=============================================================================== MESSAGE 06 ===============================================================================

>From higgins@math.niu.edu Wed Sep 9 08:44:53 1992

> We have a Sun 3/280 running SunOS 4.0.3.

Mine uses 4.1.1

> The problem: we are constantly getting > "ie0: no carrier"

Constantly = about 1 every 4-8 hours?

> errors from this machine. However, no other machine > on the same ethernet segment is having any problems. > This is the first machine on the segment. > > Sun came out and replaced the CPU board (with the > ethernet controller). I've replaced the transceiver > and transceiver cable twice (with ones that worked > fine on the other machines on this segment). > > Our engineer from sun suggested it may be a software > problem or a problem in the network. Since this is the > first machine on the segment and no other machines on this > segment are experiencing problems, I doubt its in the > ethernet itself (but I could be wrong). Does anyone have > any clue as to what may be causing this problem. > > This problem is extremely important to us because this particular > machine is our primary NIS/communications/disk server.

Mine too. However I've always thought that someone somewhere is kicking a cable or something. I've gotten to the point where I ignore it if I don't get multiple indications simultaneously. Interesting to know that someone else has the same problem. Maybe I'll have to stopp beating on the users to watch where the put their feet.

-Gregory Higgins, Systems Manager, higgins@math.niu.edu

=============================================================================== MESSAGE 07 ===============================================================================

>From beyea@ERC.MsState.Edu Wed Sep 9 09:35:59 1992

Have you added any new machines to your net recently? We had a very similar problem with our 4/280 when we added a couple of 690's to the net. It seems that the ability of those 690's to spew packets was great enough to overwhelm that poor little 4/280.

_______ / | ====/ |=============================================================== | /Starkville Tim Beyea Mississippi State University | | / * | Computer Systems Administrator Engineering Research Center | | \ | Fax: (601)325-7692 2 Research Blvd. | | | | Phone: (601)325-3709 Starkville, MS | | / | beyea@ERC.MsState.Edu 39759 | ===/ |=============================================================== |______ | / | \____|

=============================================================================== MESSAGE 08 ===============================================================================

>From lemke@MITL.COM Wed Sep 9 09:37:12 1992

I've seen similar behavior before where the problem was the *AUI connector* on the cable. Many AUI connectors come with "spacers" in the prongs between the end of the prong and the cable itself. This causes the pins in the connector to actually make a pretty poor connection with the receptacles in the ethernet port on the CPU board. Despite the fact that the connection "looks" good and that the machine seems to work OK, one still gets a bunch of transceiver error messages because even when the cable is pressed up solid against the cpu board, the connection is still poor.

If you have a hardware/cable guru around, ask them if they can figure out how to remove the prongs, take out the spacers, and replace the prongs on the AUI cable. I've never done this procedure myself--our hardware guys or one of our networking gurus always did this--but I've seen it done, and as silly as it sounds, it works. Give it a try. Kennedy Lemke Network and UNIX Systems Administrator Postmaster Matsushita Information Technology Laboratory 182 Nassau Street, Third Floor Princeton, New Jersey 08542 Work Phone: (609) 497-4600 Fax Phone: (609) 497-4013 Email: lemke@MITL.COM

=============================================================================== MESSAGE 09 ===============================================================================

>From dan@geog.utoronto.ca Wed Sep 9 10:30:12 1992

I had a similar problem one time - we had a twisted pair hub with an internal terminator on the network. the internal terminator was set incorrectly to "terminated" rather than "non-terminated". the network seemed to work okay otherwise but I kept getting a "no carrier" on a sun just down the cable from the hub. perhaps you could check your terminators, and even swap out a few cables (if you can) to isolate the problem. also does your network conform to the specs (overall length and length between nodes)?

Hope this gives you some ideas. Dan

-- Dan Desousa (dan@geog.utoronto.ca) (( Don't like my .sig? Department of Geography University of Toronto \\ dial 100 St. George St. #5047, Toronto, Ontario M5S 1A1. )) 1-800-DEV-NULL Voice: (416) 978-1450 Facsimile: (416) 978-6729 //

============================================================================== MESSAGE 10 ==============================================================================

>From pthomas@netcom.com Wed Sep 9 11:26:34 1992

Hi,

We had a similar problem with our sun4/470 running sunos4.1.1. The machine also displayed the "ie0 no carrier" message continually on the console, although the network seemed ok. I tried a new CPU board, new thinnet cable, etc. Eventually I discovered that the problem was caused by a Cabletron remote bridge which was connected to the same ethernet segment as the 470. Once this bridge was placed on it's own segment the messages stopped. Do you have any other networking equipment on the ethernet besides workstations?

Phil Thomas pthomas@netcom.com

=============================================================================== MESSAGE 11 ===============================================================================

>From poffen@sj.ate.slb.com Wed Sep 9 11:33:22 1992

First off, what kind of ethernet? Thick, thin, TP?

Anyway, one thing to look at is that the Sun-3's were a little incompatible with some cables. Not in pinout, but the posts for the slide lock that hold them in. The problem is that the posts are too long, and prevent the cable from seating all the way when plugged in, making flaky connections.

What I usually do is pull one or two of the washers off the posts on the cable, making it fit more snugly. Newer machines don't have this problem.

Also, if using thick ethernet, make sure the tap is clean. Put in a new one and see, or try and plug that machine in the cable that works for the others.

Russ Poffenberger DOMAIN: poffen@sj.ate.slb.com Schlumberger Technologies ATE UUCP: {uunet,decwrl,amdahl}!sjsca4!poffen 1601 Technology Drive CIS: 72401,276 San Jose, Ca. 95110 Voice: (408)437-5254 FAX: (408)437-5246

=============================================================================== MESSAGE 12 ===============================================================================

From: Art Schoenstadt <0085P%NAVPGS.BITNET@CORNELLC.cit.cornell.edu>

The FIRST thing I would do is wheel another SUN machine into that room and attach it to the SAME transceiver cable. This will almost certainly isolate the problem as to machine versus network, i.e. if the same problem persists with the new machine, then the problem is the transceiver connector, cable, etc., NOT the Sun. Of course, the converse is true if the problem goes away.

-------------------------------------------------------------------- | Art Schoenstadt 0085P@NAVPGS.BITNET | | Code MA/Zh (Math. Dept) 0085P@cc.nps.navy.mil | | Naval Postgraduate School (408)-646-2662 | | Monterey, CA 93943 "Are we having fun yet??" | --------------------------------------------------------------------

=============================================================================== MESSAGE 13 ===============================================================================

>From Perry_Hutchison.Portland@xerox.com Wed Sep 9 14:46:28 1992

> "ie0: no carrier"

The usual suspects would be the transceiver and its cable.

Is this thicknet (RG8) or thinnet (RG58) ? If thick, the connection between the transceiver and the network cable should be checked. Vampire-tap transceivers are notorious for causing wierd problems if their installation is anything less than perfect.

> This is the first machine on the segment.

So there could be a cable problem between it and the rest of the network. Is the cable jacket marked with an IEEE 802 number, or only with an RG number, or perhaps not at all? I have seen all sorts of crazy problems caused by someone installing the wrong type of cable -- RG59 (75 ohm, used for TV) and RG62 (93 ohm, used for IBM 3270 terminals) look a lot like RG58 (50 ohm, used for many things including thin Ethernet), but only 50 ohm cable will work properly. This sort of error is far less likely in a thicknet installation, although RG11 (75 ohm) could conceiveably be used by mistake in place of RG8.

Another possible culprit is the terminator -- if this is the wrong type or poorly connected there can be "dead spots" on the cable, and if your 3/280 just happens to be located at one of them you might see this sort of problem.

Another possible cable problem is that thicknet is more fragile than it looks, and is particularly prone to invisible damage from being kinked or crushed. Thinnet is less likely to suffer from this sort of thing.

Any of these problems should be fairly easy to find by having an experienced technician check the physical layer with a TDR or analyzer.

One thing which a TDR would not find: If by "first machine on the segment" you mean that it is closest to the repeater, and other machines on this segment can communicate through the repeater with other segments, it may be too close to the repeater. There is a minimum cable length between transceivers -- something like 2 meters for thinnet and 5 meters for thick if I am remembering correctly.

============================================================================== MESSAGE 14 ==============================================================================

>From aahvdl@eye.psych.umn.edu Wed Sep 9 15:02:58 1992

If the 3/280 is your FIRST machine on the segment, you might want to check for problems near the SECOND machine!

Also, if you are running ThinWire, make sure that you have a "silver bullet" terminator at some point BEYOND the BNC T-jack on your 3/280's transceiver...

============================================================================= MESSAGE 15 =============================================================================

>From price@homer.usc.edu Wed Sep 9 19:34:43 1992

No solution, just the same problem (we have it on 2 machines), one is connected through an 8-port Delni, it is also the server. We notice no real problems with these machines (except maybe one: do you get random occasional errors in file accesses from the server giving the no carrier errors? We have a rate of about 1 error per 1000 files each .25MegBy, with a position distribution that indicates it is network related.) For the no carrier errors, we have not yet replaced hardware, but we have moved the system around on the physical network and changed cables. Keith Price price@usc.edu



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:49 CDT