Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!usc!apple!altos!megadon!clp From: kemp@uiatma.atmos.uiuc.edu (John Kemp) Newsgroups: comp.unix Subject: UTP FOOBAR FOLLOWUP Message-ID: <2290@megadon.UUCP> Date: 6 Dec 90 22:30:28 GMT Sender: clp@megadon.UUCP Organization: University of Illinois at Urbana Lines: 85 Approved: clp@megadon.UUCP Originator: kemp@uiatma.atmos.uiuc.edu Apparently-To: comp-unix@ux1.cso.uiuc.edu and the probable solution. At this point, there are indications that the problem has been corrected. Problems: 1) campus gateway sending out "giant packets" 2) IBM 320 network software choking and dying on 1) 3) DEC vs2000-uVAXII vaxcluster dropping connections 4) PLEXCOMM utp repeater showing high number of collisions A campus networking analyst hooked a lanalyzer up to our net to examine the problem. It appears that the gateway was sending out RIP packets in extremely large ethernet frames (16Kbytes). And there were also a lot of "broken" packets being sent out. This was evidenced by a high number of CRC errors and short packets. Solutions: 1) uVAXII moved off thick cable, to repeater using AUI-UTP 2) DEC transceivers and thick backbone cable junked *3) ODS 234 FOT replaced with Cabletron FOT 4) gateway updated to current software and hardware revs Our analyst thinks the ODS FOT (fiber optic transceiver) was the most likely cause of the weirdness. The ODS operated as a "host" connected to our thick cable. The Cabletron operates somewhat differently and connects straight into the AUI port on the repeater. Personally, I tend to think that the DEC transceivers didn't help things any either. We have had trouble connecting things to them in the past. And of course, the IBM software choking and dying is an obvious software deficiency. I would hope that IBM fixes this in their next release of AIX. Just send a 16KB packet to your RS6000 and see what happens. Below are diagrams of before and after configuration. I hope this information is helpful to someone else. The analysts conclusion so far is that UTP is not the problem. Questions and comments welcome... -------- john kemp ( ( )_ internet - kemp@uiatma.atmos.uiuc.edu ----- ( ( __) decnet - uiatmb::kemp --- univ of illinois (_ ( __) bitnet - {uunet,convex} -- dept of atmos sci .(____). !uiucuxc!uiatma!kemp - 105 s gregory ave ... phone - (217) 333-6881 - urbana, il 61801 ... fax - (217) 444-4393 Glossary: ... fiber --- thin === thick ~~~ UTP Tx -- DEC AUI transceiver Ax -- TP-AUI transceiver Bx -- TP-BNC transceiver UTPR -- Plexcomm multiport UTP repeater ODS -- Fiber Transceiver FIOL -- Cabletron Fiber Link xxx -- more computers ------------------------- BEFORE ------------------------------------- Gway ... ODS ===== ===== uvaxII Tx ====== Tx ===== Tx = UTPR ~~~ Ax === ibm320 ~~~ Bx --- sun --- vs2000 ~~~ ~~~ ~~~ ~~~ ~~~ Ax === xxx ~~~ Bx --- xxx ~~~ Bx --- xxx --- xxx ~~~ Bx --- xxx --- xxx --- xxx ------------------------- AFTER ----------------------------- Gway ... FIOL ====== UTPR ~~~ Ax === uvaxII ~~~ Ax === ibm320 ~~~ Bx --- sun --- vs2000 ~~~ ~~~ ~~~ ~~~ Ax === xxx ~~~ Bx --- xxx ~~~ Bx --- xxx --- xxx ~~~ Bx --- xxx --- xxx --- xxx ------------------------- END ------------------------------------- Brought to you by Super Global Mega Corp .com