Results 1 to 10 of 10
  1. #1
    Join Date
    Apr 2009
    Location
    Spain
    Posts
    130

    question about MTR result

    Hi,
    I have run MTR in one of my server, it got the result as below,
    the packet loss in node 5 is normal or not?

    Will it affect the network quality? seems there is no packet loss in destination ip, why will this happen?

    Host Loss% Last Avg Best Wrst StDev
    1. 10.17.218.142 0.0% 0.1 0.1 0.0 0.2 0.0
    2. midphase.com 0.0% 0.3 5.2 0.3 210.5 25.4
    3. po104.dar02.dal01.dallas-datacenter.com 0.0% 0.3 2.7 0.3 137.1 16.1
    4. po2.cer01.dal01.dallas-datacenter.com 0.0% 0.5 2.6 0.4 164.7 15.7
    5. xe-4-3.r03.dllstx09.us.bb.gin.ntt.net 14.2% 0.7 12.0 0.4 208.2 33.5
    6. po-4.r02.dllstx09.us.bb.gin.ntt.net 8.9% 41.8 19.9 0.5 203.5 45.4
    7. 4.68.63.221 3.3% 0.8 2.4 0.6 62.5 8.0
    8. 4.69.145.116 0.0% 6.9 5.9 0.9 14.4 3.9

  2. #2
    Join Date
    May 2006
    Location
    NJ, USA
    Posts
    6,457
    Quote Originally Posted by pedro2010 View Post
    Hi,
    I have run MTR in one of my server, it got the result as below,
    the packet loss in node 5 is normal or not?

    Will it affect the network quality? seems there is no packet loss in destination ip, why will this happen?

    Host Loss% Last Avg Best Wrst StDev
    1. 10.17.218.142 0.0% 0.1 0.1 0.0 0.2 0.0
    2. midphase.com 0.0% 0.3 5.2 0.3 210.5 25.4
    3. po104.dar02.dal01.dallas-datacenter.com 0.0% 0.3 2.7 0.3 137.1 16.1
    4. po2.cer01.dal01.dallas-datacenter.com 0.0% 0.5 2.6 0.4 164.7 15.7
    5. xe-4-3.r03.dllstx09.us.bb.gin.ntt.net 14.2% 0.7 12.0 0.4 208.2 33.5
    6. po-4.r02.dllstx09.us.bb.gin.ntt.net 8.9% 41.8 19.9 0.5 203.5 45.4
    7. 4.68.63.221 3.3% 0.8 2.4 0.6 62.5 8.0
    8. 4.69.145.116 0.0% 6.9 5.9 0.9 14.4 3.9

    It shouldn't be happening - I'd contact the datacenter and tip them off on it.
    simplywww: directadmin and cpanel hosting that will rock your socks
    Need some work done in a datacenter in the NYC area? NYC Remote Hands can do it.

    Follow my "deals" Twitter for hardware specials.. @dougysdeals

  3. #3
    Join Date
    Jun 2009
    Location
    Seattle, WA
    Posts
    506
    Contact you Datacenter, ask them regarding it.

  4. #4
    Join Date
    Apr 2009
    Location
    Spain
    Posts
    130
    Thanks, I have contacted the DC regarding this.

    and what should it be?
    no packet loss at any node?
    and about ping,
    From that result, the Wrst ping always be more than 100ms, or 200ms,

    Is this normal?

  5. #5
    Join Date
    May 2006
    Location
    NJ, USA
    Posts
    6,457
    Quote Originally Posted by pedro2010 View Post
    Thanks, I have contacted the DC regarding this.

    and what should it be?
    no packet loss at any node?
    and about ping,
    From that result, the Wrst ping always be more than 100ms, or 200ms,

    Is this normal?
    Ideally, there would be no packet loss. Here's an MTR to softlayer.com, from localhost.

    Quote Originally Posted by mtr
    [email protected]:~$ mtr -c 25 -r softlayer.com
    HOST: debian Loss% Snt Last Avg Best Wrst StDev
    1. 192.168.1.1 0.0% 25 0.5 0.5 0.4 1.2 0.2
    2. 10.68.0.1 0.0% 25 6.7 7.6 5.6 13.5 1.7
    3. dstswr2-vlan4.rh.brfdnj.cv.n 0.0% 25 6.2 8.0 6.2 31.2 4.9
    4. ??? 100.0 25 0.0 0.0 0.0 0.0 0.0
    5. ??? 100.0 25 0.0 0.0 0.0 0.0 0.0
    6. ??? 100.0 25 0.0 0.0 0.0 0.0 0.0
    7. ??? 100.0 25 0.0 0.0 0.0 0.0 0.0
    8. te1-2.cer01.wdc01.washington 0.0% 25 17.3 28.0 14.9 194.4 38.6
    9. te9-6.dar01.dal01.dallas-dat 0.0% 25 49.6 50.3 49.2 52.2 0.8
    10. po1.slr01.dal01.dallas-datac 0.0% 25 50.7 67.6 48.7 164.8 37.2
    11. www.softlayer.com 0.0% 25 50.8 50.9 49.9 52.4 0.8

    The 100% packetloss in the middle is a weird policy my ISP has, traces to anywhere are like that.
    simplywww: directadmin and cpanel hosting that will rock your socks
    Need some work done in a datacenter in the NYC area? NYC Remote Hands can do it.

    Follow my "deals" Twitter for hardware specials.. @dougysdeals

  6. #6
    Join Date
    Aug 2006
    Location
    Ashburn VA, San Diego CA
    Posts
    4,571
    packet loss on one or two hops in a route is in fact normal (no matter what anyone tells you). It's a router giving pings lower priority...NOT necessarily dropping normal traffic.

    If you see packet loss at every hop past a suspected bad hop, then you could be seeing a real problem. THIS IS THE KEY!
    Fast Serv Networks, LLC | AS29889 | Fully Managed Cloud, Streaming, Dedicated Servers, Colo by-the-U
    Since 2003 - Ashburn VA + San Diego CA Datacenters

  7. #7
    Join Date
    Aug 2006
    Location
    Ashburn VA, San Diego CA
    Posts
    4,571
    Quote Originally Posted by Dougy View Post
    Ideally, there would be no packet loss. Here's an MTR to softlayer.com, from localhost.




    The 100% packetloss in the middle is a weird policy my ISP has, traces to anywhere are like that.
    It just means they turned off ICMP echo on their routers instead of rate limit like everyone else. Probably to keep customers from thinking something is wrong when there really isn't.

    but you can see more meaning in my previous post now -- just because a couple hops show packet loss whether it be 1% or 100% doesn't mean the route is bad.
    Fast Serv Networks, LLC | AS29889 | Fully Managed Cloud, Streaming, Dedicated Servers, Colo by-the-U
    Since 2003 - Ashburn VA + San Diego CA Datacenters

  8. #8
    Join Date
    May 2005
    Location
    New York
    Posts
    736
    It's a router giving pings lower priority
    That is slightly inaccurate to say - as the pings it is passing and responding to are identical.

    What is happening is a delay or failure in performing an administrativ response (TTL Expired response, not an ICMP Echo) - With cisco routers, this is very common also when they utiise dual cpus (ASICs) where one is dedicated to data/forwarding and the other for management & administrative responses etc - which may be under load.

    Remember - a traceroute addresses all pings to the same destination - it just increases the TTL from 1 to 30 (default maximum) on each set of 3 pings. A router automatically reduces the TTL field in a packet by 1 when it forwards it. When a router sees it reach 0, and is trying to forward the packet, it responds with TTL expired and discards the ping packet. When it reaches the addressed destination, it will correctly respond with the echo signalling the end of the traceroute.

    If a router is configured not to reespond to pings, or not sent TTL Expires - then they would not show up in a traceroute (would show as * * * and be marked as a hop, as it still reduces the TTL and forwards the packet)
    Last edited by jNive; 07-13-2009 at 11:37 PM.
    Perigee Global Corporation
    Design, Development and Hosting Solutions
    Dedicated Servers, CDN, Hosted E-Mail, Web Hosting, VPS & Cloud Servers
    1.212.400-7632 www.perigeeglobal.com

  9. #9
    Join Date
    Oct 2002
    Location
    Vancouver, B.C.
    Posts
    2,656
    Quote Originally Posted by jNive View Post
    That is slightly inaccurate to say - as the pings it is passing and responding to are identical.

    What is happening is a delay or failure in performing an administrativ response (TTL Expired response, not an ICMP Echo) - With cisco routers, this is very common also when they utiise dual cpus (ASICs) where one is dedicated to data/forwarding and the other for management & administrative responses etc - which may be under load.
    That's not quite accurate either. A router's response involves both receiving the packet, as well as sending the response packet back. In some cases, such as with Juniper and Foundry routers, there is a hard limit on ICMP per interface. This will still have an impact on incoming packets on traceroutes, regardless of the function of the response. Also, a TTL expired response still uses the ICMP protocol, be it for ICMP, UDP, or TCP based traceroutes. You are correct however, in that it is not an echo response.

    ASICs should not be thought of as CPU's as they are the exact opposite of general purpose processors. ASICs are specialized for a specific function. There can be one or many line cards in a router, and each line card can have one or many ASICs. This architecture is completely different than a desktop computer with 'dual cpus', and should not be described as such. Whereas responses to traceroutes are performed by the main CPU, the function of forwarding packets can be performed solely by line cards without the main CPU. Thus, the main CPU can be fully occupied by BGP updates or whatnot, resulting in slow responses to traceroute packets without having any impact on passing traffic. The explanation of low icmp (though not ping) priority in this case is still correct; it is just a version of the above that has been simplified to be more palatable to the general customer.

    Anyway, I would concur with FastServ that packet loss isolated to a single hop is no cause for concern whatsoever. It is only when packet loss persists through all further hops that there is true packet loss occurring.
    ASTUTE HOSTING: Advanced, customized, and scalable solutions with AS54527 Premium Canadian Optimized Network (Level3, PEER1, Shaw, Tinet)
    MicroServers.io: Enterprise Dedicated Hardware with IPMI at VPS-like Prices using AS63213 Affordable Bandwidth (Cogent, HE, Tinet)
    Dedicated Hosting, Colo, Bandwidth, and Fiber out of Vancouver, Seattle, LA, Toronto, NYC, and Miami

  10. #10
    Join Date
    May 2005
    Location
    New York
    Posts
    736
    yeah - the term cpu was used pretty loosely :-) and while the response of course involves receiving + response, I was talking primarily in relation to cases where a response from a router is not received or is delayed in a traceroute. Other than that I think everyone is on the same page hehe!!
    Perigee Global Corporation
    Design, Development and Hosting Solutions
    Dedicated Servers, CDN, Hosted E-Mail, Web Hosting, VPS & Cloud Servers
    1.212.400-7632 www.perigeeglobal.com

Similar Threads

  1. Simple question regarding sorting SQL Result
    By latheesan in forum Programming Discussion
    Replies: 1
    Last Post: 06-28-2008, 02:26 PM
  2. mtr result help. (result attached)
    By SICStus in forum Dedicated Server
    Replies: 3
    Last Post: 02-23-2008, 09:16 AM
  3. Result.
    By phill2003 in forum Web Hosting Lounge
    Replies: 1
    Last Post: 08-03-2006, 02:43 PM
  4. How can i get several PTR result
    By imer_cleric in forum Hosting Security and Technology
    Replies: 3
    Last Post: 03-17-2004, 10:14 AM
  5. Java Result Set Question
    By cosmic_bird in forum Programming Discussion
    Replies: 10
    Last Post: 03-28-2003, 12:00 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •