dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
1793
share rss forum feed


bluepoint

join:2001-03-24

1 edit

Bad Gateway

Since power came back in our area after Sandy, our gateway is not responding to traceroute even if I set the timeout @10000ms. I'm not complaining yet, I know Sandy did some damage in our area but at some point if this doesn't improve I'll call support and see what they got.

Host Information for "myservices.timwewarnercable.com":
  IP address:   69.43.161.162
  Host name:    myservices.timwewarnercable.com
 
Tracing route to "69.43.161.162" over max. 30 hops:
 
   1:     0 ms  192.168.1.1
   2:     *     Request timed out.
   3:    15 ms  24.164.129.218  (24-164-129-218.si.rr.com)
   4:    15 ms  24.164.128.34   (tenge-0-4-0-4-nycmny1-rtr02.si.rr.com)
   5:    15 ms  184.152.112.147 (bun119.nycmnytg-rtr001.nyc.rr.com)
   6:    15 ms  24.29.148.250   (bun6-nycmnytg-rtr002.nyc.rr.com)
   7:    31 ms  107.14.19.24   
   8:    15 ms  66.109.6.161    (ae-1-0.pr0.nyc30.tbone.rr.com)
   9:    15 ms  216.55.0.65    
  10:    77 ms  216.156.0.21    (vb1011.rar3.washington-dc.us.xo.net)
  11:    78 ms  207.88.12.9     (te-3-0-0.rar3.atlanta-ga.us.xo.net)
  12:    78 ms  207.88.12.2     (te-3-0-0.rar3.dallas-tx.us.xo.net)
  13:    78 ms  216.156.0.126   (ae0d0.mcr2.sandiego-ca.us.xo.net)
  14:    93 ms  216.55.27.26   
  15:   155 ms  69.43.169.112   (ge5-2-6509-a.castleaccess.com)
  16:    93 ms  69.43.129.188  
  17:    93 ms  69.43.161.162  
 
The target host "69.43.161.162" have been reached.
 
Query finished.
 


DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15

The gateway is set to not reply to ICMP packets, which is a common configuration. Nothing to improve since it's not broken and not hurting your connection. It's certainly not "bad".



Napsterbater
Meh
Premium,MVM
join:2002-12-28
Milledgeville, GA
reply to bluepoint

What are you gonna complain about? nothing is wrong.



bluepoint

join:2001-03-24
reply to DrDrew

It was responding before Sandy, did they do the change?



bluepoint

join:2001-03-24
reply to Napsterbater

How did you know nothing is wrong?



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15
reply to bluepoint

said by bluepoint:

It was responding before Sandy, did they do the change?

Apparently they did. Although I've seen the same behavior if something on the home network is blocking RFC 1918 (private) addressed packets from crossing the WAN port and/or some VPN setups.

What exactly do you see wrong besides a single hop not responding to ICMP packets?


bluepoint

join:2001-03-24

1 edit

My son is complaining of slow gaming on his ps3 and without knowing the gateways performance I cannot figure out the problem. Is there any reason you guys choose to do this?
Oh BTW, you said apparently, meaning you're not sure?



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15

I didn't choose to do anything, I'm not involved. Local TWC CMTS Engineers may have done it.

Figure the gateways performance is equal to the hop right after it, since all the traffic hitting the next hop has to go through the gateway. If the gateway hop was really causing a problem, you'd see it reflected in all of the hops after it.

I said "apparently" because I only have what you've posted to judge the issue on. It may be a couple of other reasons as I listed though.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.



bluepoint

join:2001-03-24

Okay, so seems you're not sure if it was done. I asked because two days ago I've seen it answer with 2000+ ms, an indication of a busy gateway.
Sometimes you can gauge the gateway's performance through the next hop to it, however, it's not always true. I've seen a very long response from the gateway but very short response from the next hop after it.

If they do so, how are they going to diagnose those routers? There is really no advantage for TWC to have those gateways not respond to ICMP packets.



Jabbu
Premium
join:2002-03-06
reply to bluepoint

Ping traffic is lowest priority, if you get a 2000ms in tracert, but the final hop is only 50ms, it means your data round trip to target is only 50ms, the 2000ms is not the issue, if it was a real issue you would see 2000ms, then 2005ms, 2011ms, 2016ms for example all the way to your target at lets say 2060ms.

Ping uses ICMP, and ICMP is the LAST thing a router may care about, hence your tracert showing a really high hop.



bluepoint

join:2001-03-24

I do not agree, each hop contributes to latency and latency is important when you're gaming.



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15
reply to bluepoint

said by bluepoint:

Okay, so seems you're not sure if it was done. I asked because two days ago I've seen it answer with 2000+ ms, an indication of a busy gateway.
Sometimes you can gauge the gateway's performance through the next hop to it, however, it's not always true. I've seen a very long response from the gateway but very short response from the next hop after it.

That doesn't mean the gateway is busy. It means the card management processor may be configured to put ICMP on a lower priority or the card manager is busy, but that frequently doesn't affect the packet passthrough on the gateway to another router.

Also, 2000+ms response time could be from your own network slowing down pings to the hops after it.

Again, though it could be private addresses are blocked at the WAN port of your home router causing the no response.

Lots of possibilities.... especially after a wide spread power outage.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15
reply to bluepoint

said by bluepoint:

I do not agree, each hop contributes to latency and latency is important when you're gaming.

If the gateway hop was contributing to the latency in an unusual manor, you'd see it reflected in the hops following it.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


Jabbu
Premium
join:2002-03-06
reply to bluepoint

said by bluepoint:

I do not agree, each hop contributes to latency and latency is important when you're gaming.

What do you not agree with?

Ping is ICMP, ICMP traffic routers do not care about, you know about QoS or CoS?

How do you explain seeing 2000ms on your first hop, but only 50ms on your destination?


bluepoint

join:2001-03-24
reply to DrDrew

said by DrDrew:

That doesn't mean the gateway is busy. It means the card management processor may be configured to put ICMP on a lower priority or the card manager is busy, but that frequently doesn't affect the packet passthrough on the gateway to another router.

So, it's busy tending to other higher priority requests. Before Sandy, the gateway answers within 30 ms anytime of day. How can you explain that?

Again, though it could be private addresses are blocked at the WAN port of your home router.

Like what I've said, the gateway has been responding everytime before the storm and as far as my network, the gateway is always seen as a public ip(68.xx.xx.xx).


bluepoint

join:2001-03-24
reply to Jabbu

said by Jabbu:

said by bluepoint:

I do not agree, each hop contributes to latency and latency is important when you're gaming.

What do you not agree with?

Ping is ICMP, ICMP traffic routers do not care about, you know about QoS or CoS?

How do you explain seeing 2000ms on your first hop, but only 50ms on your destination?

So you are telling me ICMP packets are ignored and it means we do not pay attention to the trace results? The ms it represents for each hop is the speed the hop answers. It doesn't represent how fast it arrives to the hop. It's wrong to ignore the other routes and just be happy the the destination answers to your satisfaction.
In layman's words, it's just the same in a car race. One starts in a rough road and the other in a smooth road. If both cars races to reach the 1 mile finish line, which one do you think arrives first?


DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15
reply to bluepoint

So before Sandy it was 30ms, now it's not responding? In between there was a large power outage. Patterns of usage have changed and software may have automatically been updated once the power came up.

During large power outages, IPs maybe shuffled and reassigned.

There are far too many things that can change to rely on a single posted trace route to really troubleshoot. All we're doing right now is guessing.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.



Jabbu
Premium
join:2002-03-06
reply to bluepoint

I don't get your point.

You are saying your first hop is taking 2000ms, that would mean ALL your websites would take at least 2 seconds to start loading. All your gaming would have a 2000ms or more ping time. If this was real, a simple ping to google.com would also result in 2000ms, not 60ms.

The first hop is the router in the hub, its priority is phone traffic then data traffic and its least priority is icmp traffic. So if this router is busy, you WILL see a high delay in tracert.



bluepoint

join:2001-03-24

2 edits
reply to DrDrew

Ip's being reshuffled didn't happen at least here. We lost power for 3.5 days and I was surprise when the power came back the DHCP server gave me the same IP. It's like winning the lottery.

All you're saying are guesses, I want a factual answer and I posted to get an answer and not to argue. I want to know why the gateway used to ping 30ms and now answering @2000ms or not at all.



bluepoint

join:2001-03-24

1 edit
reply to Jabbu

said by Jabbu:

The first hop is the router in the hub, its priority is phone traffic then data traffic and its least priority is icmp traffic. So if this router is busy, you WILL see a high delay in tracert.

That's what I'm trying to point at the router is busy. Why was it pinging 30ms before the storm and 2000ms now? What do you think happens? Wouldn't you think gaming will be affected?

"Network latency in a packet-switched network is measured either one-way (the time from the source sending a packet to the destination receiving it), or round-trip (the one-way latency from source to destination plus the one-way latency from the destination back to the source). Round-trip latency is more often quoted, because it can be measured from a single point. Note that round trip latency excludes the amount of time that a destination system spends processing the packet. Many software platforms provide a service called ping that can be used to measure round-trip latency. Ping performs no packet processing; it merely sends a response back when it receives a packet (i.e. performs a no-op), thus it is a relatively accurate way of measuring latency."
»en.wikipedia.org/wiki/Latency_(engineering)


DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15

1 edit

said by bluepoint:

That's what I'm trying to point at the router is busy. Why was it pinging 30ms before the storm and 2000ms now? What do you think happens? Wouldn't you think gaming will be affected?

It's not 2000ms now, it's not responding now. At least that's what you said and showed originally.

If it's 2000ms then show us.

Use PingPlotter, run it for awhile, and post some graph screenshots. A single tracert with single hop times don't show enough to really give a good answer.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


bluepoint

join:2001-03-24

1 edit

Ok I just did a tracert and it's better and it's responding. :uhh:

 1:     0 ms  192.168.1.1
2:   458 ms  68.174.228.1    (cpe-68-174-228-1.si.res.rr.com)
3:    15 ms  24.164.129.218  (24-164-129-218.si.rr.com)
4:     0 ms  24.164.128.34   (tenge-0-4-0-4-nycmny1-rtr02.si.rr.com)
5:    15 ms  184.152.112.147 (bun119.nycmnytg-rtr001.nyc.rr.com)
6:    15 ms  24.29.148.250   (bun6-nycmnytg-rtr002.nyc.rr.com)
7:    15 ms  66.109.6.78     (ae-4-0.cr0.nyc30.tbone.rr.com)
8:    15 ms  107.14.19.153  
9:    31 ms  4.30.135.153    (xe-11-0-2.edge2.Newark1.Level3.net)
10:    31 ms  4.69.156.30     (ae-31-51.ebr1.Newark1.Level3.net)
11:    15 ms  4.69.132.97     (ae-2-2.ebr1.NewYork1.Level3.net)
12:    15 ms  4.69.134.66     (ae-61-61.csw1.NewYork1.Level3.net)
13:    15 ms  4.69.155.6      (ae-14-60.car4.NewYork1.Level3.net)
14:    16 ms  4.53.94.218     (SPLICE-COMM.car4.NewYork1.Level3.net)
15:    15 ms  67.215.65.132   (hit-nxdomain.opendns.com)
 
The target host "67.215.65.132" have been reached.
 
Query finished.
 
   1:     0 ms  192.168.1.1
   2:   405 ms  68.174.228.1    (cpe-68-174-228-1.si.res.rr.com)
   3:     0 ms  24.164.129.218  (24-164-129-218.si.rr.com)
   4:    15 ms  24.164.128.34   (tenge-0-4-0-4-nycmny1-rtr02.si.rr.com)
   5:    31 ms  184.152.112.147 (bun119.nycmnytg-rtr001.nyc.rr.com)
   6:    15 ms  24.29.148.250   (bun6-nycmnytg-rtr002.nyc.rr.com)
   7:    16 ms  107.14.19.24   
   8:    15 ms  66.109.6.161    (ae-1-0.pr0.nyc30.tbone.rr.com)
   9:    15 ms  216.55.0.65    
  10:    98 ms  216.156.0.21    (vb1011.rar3.washington-dc.us.xo.net)
  11:    77 ms  207.88.12.9     (te-3-0-0.rar3.atlanta-ga.us.xo.net)
  12:    94 ms  207.88.12.2     (te-3-0-0.rar3.dallas-tx.us.xo.net)
  13:    78 ms  216.156.0.126   (ae0d0.mcr2.sandiego-ca.us.xo.net)
  14:   358 ms  216.55.27.26   
  15:    94 ms  69.43.169.112   (ge5-2-6509-a.castleaccess.com)
  16:   109 ms  69.43.129.188  
  17:    93 ms  69.43.161.162  
 
The target host "69.43.161.162" have been reached.
 
Query finished.
 

I think they are probably working on those nodes.


Jabbu
Premium
join:2002-03-06
reply to bluepoint

What are you seeing? Constant lag or spikes?



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15
reply to bluepoint

Read this:
»www.nessoft.com/kb/article/packe···-24.html
or this:
»www.nessoft.com/kb/2



bluepoint

join:2001-03-24
reply to Jabbu

Gaming is lagging. My son always got shot easily.



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15

said by bluepoint:

Gaming is lagging. My son always got shot easily.

What game and server is he connected to? How? PC, console?


bluepoint

join:2001-03-24

It's a PS3 and I have to ask my son which server. I would think the server is here in NYC. My last post, the first trace, is to ps3.sony.com.



DrDrew
That others may surf
Premium
join:2009-01-28
SoCal
kudos:15

said by bluepoint:

I would think the server is here in NYC. My last post, the first trace, is to ps3.sony.com.

ps3.sony.com could be totally unrelated to the server your son uses. It's like having some one trace to www.twcable.com when your own IP has issues.

Traces to the actual IP your son uses would be MUCH more helpful.
--
If it's important, back it up... twice. Even 99.999% availability isn't enough sometimes.


bluepoint

join:2001-03-24

Okay then I have to wait, he won't be here until later.



Jabbu
Premium
join:2002-03-06
reply to bluepoint

I'm sure there are a lot of issues in NYC area with servers and routers.