no ip mroute-cache, help plz!

tech671tech671 Member Posts: 17 ■□□□□□□□□□
Hi all,

I just received my CCNA and aspiring to be CCNP.
I have a problem at work and my boss asked me to look at our Cisco router config and I don't know if I am on the right forum to ask about ip mroute-cache issue.

Problem is our Wan connection is really slow between the two office running 2 T1 connection on a multilink. On my analysis I found the bottle neck on E0/0 on half duplex running to our LAN. I changed it to Full duplex then the E0/0 shutdown, since this was unscheduled outage I just put it back to half. BTW the E0/0 is connected to 2950 switch and settings are all auto. I have a shot on fixing this problem this weekend and be the hero on fixing our WAN connection. So I have been researching ip mroute-cache theres alot of docs and I have not yet pin point what I need. I understand ip mroute-cache is used to enable Fast switching (MDS) and it's disabled to use with multilink ppp, so far from what I am seeing.

Question: On E0/0 interface, is it right to changed the no ip mroute-cache to ip mroute-cache so that I can enable Full duplex not cause the connection between E0/0 to LAN (switch) not work? Does this solve the shutdown of the interface and hopely the bottle neck issue?

Looking for insight.
Thank you very much in advance!

Here is the config:

hostname Office2-router
!
logging buffered 4096 debugging
enable secret cisco
enable password cisco
!
username admin privilege 15
ip subnet-zero
ip cef
!
call rsvp-sync
!
interface Multilink1
bandwidth 2560
ip address 10.1.3.2 255.255.255.252
ppp multilink
multilink-group 1
!
interface Ethernet0/0
ip address 10.1.2.2 255.255.255.0
no ip mroute-cache
half-duplex
!
interface Serial0/0
description Full T-1 to Office1 Data
bandwidth 1536
no ip address
encapsulation ppp
no ip mroute-cache
service-module t1 clock source internal
ppp multilink
multilink-group 1
!
interface Ethernet0/1
no ip address
shutdown
half-duplex
!
interface Serial0/1
description Voice and Data to Office1
bandwidth 1024
no ip address
encapsulation ppp
no ip mroute-cache
ppp multilink
multilink-group 1
!
ip classless
ip route 0.0.0.0 0.0.0.0 Multilink1
ip route 0.0.0.0 0.0.0.0 10.1.1.4
ip route 10.1.1.0 255.255.255.0 Multilink1
ip route 10.1.5.0 255.255.255.0 10.1.1.90

Comments

  • mikej412mikej412 Member Posts: 10,086 ■■■■■■■■■■
    Lets's see.....

    ip mroute-cache -- do you even have multicast routing enabled? What multicast traffic do you have? Shouldn't have anything to do with half-duplex vs full-duplex.

    The old Ethernet Interface probably is half-duplex. If you want full-duplex, upgrade to FastEthernet.

    Plus 2x1.54Mbps is still less than the actual throughput of 10Mb Ethernet, so if you have local traffic from 100Mb interfaces on that 2950 routing through that router's 10Mb interface -- upgrade it. But you have Serial 1 configured for 1024 -- so its a fractional T1? So you're doing multilink with two Serial Ports that are not equal in bandwidth. Are the interface statistics showing any problems?

    service-module t1 clock source internal is set on one interface -- uh, any idea why you're not getting the clock from the line? Is that serial connection directly connected back-to-back with a T1 cross-over cable?

    You've got the comment about voice and data -- but you don't have any QOS configured. If someone has a large data transfer, your voice traffic is going to take a big hit. I don't see fragmentation and interleaving enabled -- again, Voice will get crushed by large data transfer packets.

    What does your routing table show? You have two default static routes, one using the Multilink interface, and the other using another IP. Are both of them in your routing table?

    What do pings and/or traceroutes show?
    :mike: Cisco Certifications -- Collect the Entire Set!
  • cisco_troublecisco_trouble Inactive Imported Users Posts: 78 ■■□□□□□□□□
    QoS is not a MUST when using voice, it might not be there core concern. And by default, WFQ is most likely activated on the serials anyways. Less than 2.048kb/s.
    "To strive, to seek, to find, and not to yield."
  • EdTheLadEdTheLad Member Posts: 2,111 ■■■■□□□□□□
    WFQ is not good enough for voice if the link is congested.
    Networking, sometimes i love it, mostly i hate it.Its all about the $$$$
  • tech671tech671 Member Posts: 17 ■□□□□□□□□□
    Thank you all for the responses.

    Here is the ip route output:

    10.0.0.0/8 is variably subnetted, 5 subnets, 3 masks
    C 10.1.3.0/30 is directly connected, Multilink1
    C 10.1.3.1/32 is directly connected, Multilink1
    C 10.1.2.0/24 is directly connected, Ethernet0/0
    S 10.1.1.0/24 is directly connected, Multilink1
    S 10.1.5.0/24 [1/0] via 10.1.1.90
    S* 0.0.0.0/0 is directly connected, Multilink1
    [1/0] via 10.1.1.4

    The wan link is using Adran TSU 120 via serial connection and the other is just a cat5 running from the wall to the CSU/DSU card on the router. Voice is not being used on this routes. I think they were preparing it long time ago and left the description on the interface. As far as the other info I am not sure, it's still beyond me. I have not started CCNP level yet still at the CCNA level. But I think it's a good way to get started.

    I guess my question at this point is do can I need no ip mroute-cache under Interface E0/0? Should I take it out? My main goal is change the half duplex to full. Would this help alot by changing it to full. Isn't half duplex a big bottle neck even with wan connection at 1.5 mbps then from wan to lan at half duplex 10mbps, don't you think or does it even matter since WAN is 1.5 ? please correct on this.
  • YankeeYankee Member Posts: 157
    Don't worry about the mroute cache statement as it is not causing your problem. Looks to me like you have two default routes configured and
    ip route 0.0.0.0 0.0.0.0 10.1.1.4 should be removed.

    If that is a 2611 router I seem to remember even though it had a full duplex command it did not work. If your ethernet is not taking errors other than a low collision percentage which would be normal then it is not your problem. To change the duplex, first change the speed to 10 (not auto) then change the duplex. My guess is it will not really be in full duplex (assuming this is a 2611) because it doesn't really have that functionality.

    Yankee
  • tech671tech671 Member Posts: 17 ■□□□□□□□□□
    I worked on it over the weekend. I am new to assuming the network administration work. I just found out that we are indeed using voice and data on one of the serial link. The multilink between the two office is at 2.5mbps. I able to changed the e0/0 to full duplex and the 2950 switch to Full-duplex and auto speed which is now running full at 10mb, I just realized that 2611's ethernet interface is only capable of 10 mbps. But I think I accomplished my main goal is to run on full duplex. I am not sure about enabling ip mroute-cache though I don't know if does any good or bad. What do you think?
  • YankeeYankee Member Posts: 157
    Look for late collisions on the router E0/0 interface. If they are present then E0/0 is not really in full duplex. The switch side will show CRCs and probably frame errors if the router is actually running half-duplex.

    Mroute as the name implies is used in conjunction with multicast, so if E0/0 is not PIM enabled (assuming that is your multicast protocol) then remove it. Either way I do not believe it is the cause of any of your issues.

    You did not address why there were two default routes configured.

    Yankee
  • darkuserdarkuser Member Posts: 620 ■■■□□□□□□□
    any time you disable autonegotiation you should hard code both end of the link.
    rm -rf /
  • YankeeYankee Member Posts: 157
    Agreed Darth, too many issues with auto-negotiate but I expect no replies to this thread from the originator. He's trying hard.

    Yankee
  • tech671tech671 Member Posts: 17 ■□□□□□□□□□
    Thanks Guys for the insights, I apologize for the delayed response, got caught up with other stuff, but any way. It's running on 10mb, I will change the switch to 10mb. There is no CRC error on the switch but CRC error occurs on the e0/0 interface. The 10.1.1.4 is a PIX appliance. I have not really seen a big diffence in WAN speed. I think the real problem is more on application congestions, we have a several wan application traveling between the two office, we are in the process of eleminating some and rethinking our network engineering and also upgrading our cisco routers which requires prudent planning. I thought by running full will significantly changed the speed or at least remove the Windows Exchange requesting for data widow which irritates everyone. I wish I would have taken a baseline before and after. I will before we change or upgrade things around.


    Ethernet0/0 is up, line protocol is up
    Hardware is AmdP2, address is 00d0.5839.41c0 (bia 00d0.5839.41c0)
    Internet address is 10.1.2.2/24
    MTU 1500 bytes, BW 10000 Kbit, DLY 1000 usec,
    reliability 255/255, txload 1/255, rxload 1/255
    Encapsulation ARPA, loopback not set
    Keepalive set (10 sec)
    ARP type: ARPA, ARP Timeout 04:00:00
    Last input 00:00:00, output 00:00:00, output hang never
    Last clearing of "show interface" counters never
    Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0
    Queueing strategy: fifo
    Output queue: 0/40 (size/max)
    5 minute input rate 22000 bits/sec, 8 packets/sec
    5 minute output rate 63000 bits/sec, 6 packets/sec
    2721910 packets input, 660194053 bytes, 0 no buffer
    Received 615915 broadcasts, 0 runts, 0 giants, 0 throttles
    9 input errors, 6 CRC, 2 frame, 0 overrun, 3 ignored
    0 input packets with dribble condition detected
    2406140 packets output, 1449278035 bytes, 0 underruns
    268 output errors, 0 collisions, 2 interface resets
    0 babbles, 0 late collision, 0 deferred
    268 lost carrier, 0 no carrier
    0 output buffer failures, 0 output buffers swapped out
Sign In or Register to comment.