[j-nsp] Half Duplex Routing Engine m40?

Wed Feb 16 14:41:04 EST 2005

In both cases, the duplex of the interface on the RE is the
result of negotiation with "the other end".  The M40 SCB is
only a half-duply interface, so regardless of the RE's
capability, the fxp1 can only run in half-dup mode.  On the
other hand, if your fxp0 "management" Ethernet port is
attached to a full-duplex device, then it will run run in
full-dup mode.

-----Original Message-----
From: Dan Benson [mailto:dbenson at swingpad.com]
Sent: Wednesday, February 16, 2005 11:37 AM
To: Paul Goyette
Cc: Dennis Ponne; juniper-nsp at puck.nether.net
Subject: Re: [j-nsp] Half Duplex Routing Engine m40?

 From what I know, there are two FXPs, one to the Craft interface on the
front of the unit, and the other connects to the SCB.  On all of my
M40's, I have full to the craft and half to the SCB.  Hope this helps.. //db

Paul Goyette wrote:

>The M40 internal PFE-to-RE link (fxp1) runs only in
>half-duplex mode.
>
>-----Original Message-----
>From: juniper-nsp-bounces at puck.nether.net
>[mailto:juniper-nsp-bounces at puck.nether.net]On Behalf Of Dennis Ponne
>Sent: Wednesday, February 16, 2005 11:15 AM
>To: juniper-nsp at puck.nether.net
>Cc: noc at netholding.nl
>Subject: [j-nsp] Half Duplex Routing Engine m40?
>
>
>Hi,
>
>I found something what I think is not correct.
>
>The routing-engine on our M20 box on FXP1 has a full-duplex link.
>
>dennis at m20.intxn.be.as24730.net> show interfaces fxp1
>Physical interface: fxp1, Enabled, Physical link is Up
>  Interface index: 2, SNMP ifIndex: 2
>  Type: Ethernet, Link-level type: Ethernet, MTU: 1514, Speed: 100mbps
>  Device flags   : Present Running
>  Interface flags: SNMP-Traps
>  Link type      : Full-Duplex
>  Link flags     : 4
>  Current address: 00:a0:a5:12:1c:1e, Hardware address: 00:a0:a5:12:1c:1e
>  Last flapped   : Never
>  Input packets : 332905148
>  Output packets: 994816558
>
>
>Now I check both our M40's and check this out:
>
>Physical interface: fxp1, Enabled, Physical link is Up
>  Interface index: 2, SNMP ifIndex: 2
>  Type: Ethernet, Link-level type: Ethernet, MTU: 1514, Speed: 100mbps
>  Device flags   : Present Running
>  Interface flags: SNMP-Traps
>  Link type      : Half-Duplex
>  Link flags     : 4
>  Current address: 00:a0:a5:12:2b:53, Hardware address: 00:a0:a5:12:2b:53
>  Last flapped   : Never
>  Input packets : 1643478
>  Output packets: 5153017
>
>
>What we have we do a total on this M40 box 2.9 - 3.1 gigabit/s of traffic
>outgoing and incomming 600 - 700 mbit/s on our transit/peering interfaces.
>
>We run 197 peers on this box and the avaible memory is 351MB, this box has
>768MB of ram inside and uses a RE-2.
>
>The total pps is:
>
>250.000kpps average, Load in day time 300.000kpps Load up to 320.000kpps.
>
>But the strange is that we have packet loss, it does not matter wich JunOS
>we use, the packet loss to the router is heavy, but 9 of 10 times after
>the router the ping is normal. But on the M20 this is not an issue.
>
>How can I force the FXP1 to be 100 FDX? Threw the software:
>
>edit interface fxp1
>set link-mode full-duplex
>
>Does not work.
>
>JUNOS Base OS Software Suite [7.0R2.7]
>
>Hardware inventory:
>Item             Version  Part number  Serial number     Description
>Chassis                                53828             M40
>Backplane        REV 02   710-001348   AL3546            M40 Backplane
>Power Supply A   REV 04   740-000234   001516            AC Power Supply
>Power Supply B   REV 04   740-000234   001598            AC Power Supply
>Maxicab          REV 08   710-000229   AT3658            Host Maxi CAB
>Minicab          REV 03   710-001739   AH8902            M40 Teknor Mini
>CAB
>Display          REV 08   710-000150   AP2907            M40 Display Board
>Routing Engine   REV 01   740-003239   9001009622        RE-2.0
>SCB              REV 02   710-001838   AP9305            Internet
>Processor IIv1
>FPC 0            REV 01   710-001292   AN7304            FPC
>  PIC 0          REV 03   750-002879   AN7060            1x G/E, 1000
>BASE-SX
>  PIC 1          REV 03   750-002879   AN7060            1x G/E, 1000
>BASE-SX
>  PIC 2          REV 03   750-002879   AN7060            1x G/E, 1000
>BASE-SX
>  PIC 3          REV 03   750-002879   AN7060            1x G/E, 1000
>BASE-SX
>FPC 2            REV 03   710-003308   BC0734            E-FPC
>  PIC 0          REV 01   750-001324   AC4850            1x G/E, 1000
>BASE-LX
>FPC 5            REV 01   710-001292   AF3547            FPC
>  PIC 0          REV 03   750-002879   AP1259            1x G/E, 1000
>BASE-SX
>  PIC 1          REV 03   750-002879   AP1259            1x G/E, 1000
>BASE-SX
>  PIC 2          REV 03   750-002879   AP1259            1x G/E, 1000
>BASE-SX
>  PIC 3          REV 03   750-002879   AP1259            1x G/E, 1000
>BASE-SX
>FPC 7            REV 01   710-001292   AJ0968            FPC
>  PIC 0          REV 03   750-000603   AG7521            4x OC-3 SONET,
>SMIR
>  PIC 2          REV 08   750-001072   AH1073            1x G/E, 1000
>BASE-SX
>
>dennis at m40.ams.gs> show system uptime
>Current time: 2005-02-16 20:14:55 CET
>System booted: 2005-01-26 23:09:08 CET (2w6d 21:05 ago)
>Protocols started: 2005-01-26 23:10:23 CET (2w6d 21:04 ago)
>Last configured: 2005-02-16 19:49:19 CET (00:25:36 ago) by dennis
>8:14PM CET up 20 days, 21:06, 1 user, load averages: 0.02, 0.14, 0.22
>
>The question is how can make sure that a router that is no heavy loaded
>towards it's routing engine can ping back in time? How do I set this
>interface to be full-duplex?
>
>Thanks in advance,
>
>Dennis Ponne
>
>_______________________________________________
>juniper-nsp mailing list juniper-nsp at puck.nether.net
>http://puck.nether.net/mailman/listinfo/juniper-nsp
>
>_______________________________________________
>juniper-nsp mailing list juniper-nsp at puck.nether.net
>http://puck.nether.net/mailman/listinfo/juniper-nsp
>
>
>
>