cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1452
Views
75
Helpful
13
Replies

BGP up Time and Late collision on WAN interface

MrBeginner
Spotlight
Spotlight

Dear Sir,

I would like to request to explain about BGP up/downtime and collision late error.please?

when i check using show ip bgp summary,we can see as below .In the Up/Down is show BGP uptime?

Neighbor        V    AS MsgRcvd MsgSent   TblVer  InQ OutQ Up/Down State/PfxRcd
10.100.1.1      4   200      26      22      199    0    0 00:14:23 23
10.200.1.1      4   300      21      51      199    0    0 00:13:40 0

 I always saw this time is 5week3days 6weeks1days..etc.But i didn't see 1 years or 2 years.Let me know BGP reset their UP time ? Let me know what time will reset this up time ?

 

When i check in my router i saw LATE collision error message in our WAN link.Let me know it can be effect to our internet speed ?

1 Accepted Solution

Accepted Solutions

Hello MrBeginner,

you have not cleared the counters.

We look at the output section

389785200 packets output

6176889 output errors,

 

the error ratio is calculated as : errored packets / (output packets + errored packets)=

= 6176889 / (6176889+389785200) = 0,01559969798017708710492230987295 = 1,559 %

 

However, to be noted there is an high number of interface resets :

>> 1463886 interface resets

This are not good also the late collisions if increasing should mean the other side is working in full duplex.

 

So check the LAN cable, change it with one that you know works well and try to use static settings

of

speed 100

duplex full

 

clear the counters on the interface with:

clear counters giga9

 

Note:

the use of speed 100 duplex full may solve or it can make the performance worse. Check with the ISP if their router supports 100 Mbps full duplex. If ISP router is able the issue should be on the LAN cable.

So changing the LAN cable is the first wise action to do.

 

Hope to help

Giuseppe

 

View solution in original post

13 Replies 13

Mark Malone
VIP Alumni
VIP Alumni
Hi
Its stating link is only up this amount of time ,so it failed 0 00:14:23 23 , 14 minutes ago and is now back up with 23 BGP routes available , your show logs should confirm 6the times too

please post the output of the show interface for the wan link so we can see exactly what errors are on it
has it always been like this , your settings may be incorrect or if its only happened you may have a WAN issue

Hi,

Thanks for your help. Can I request to explain about BGP up time?

I want to if the neighbor router is no issue, bgp is always up?May be one year,2year ,3year? When they reestablished?

 

The late collision message is got other issue.It is ISP issue.because my site is auto config. I changed half duplex .still got error.ISP work half dulplex.I would like to know this Late collision error will case to slow internet speed or we can get more delay?

Yes in theory it could stay up that long no problem but in the real world at some stage the ISP will introduce new hardware or do a software upgrade which would take down the circuit and that would cause the timers to reset ,whenever the BGP neighbour goes down ,the interface goes down or the circuit itself fails these would cause outages to the BGP relationship and reset the timers


Auto connected direct to ISP is not a usual setting from my experience, try and set the speed to like full duplex speed 100 or 10 ,usually its hardcoded when connected to an ISP ,any collisions are not good in a live environment and you should try to avoid them , but if you have to go half duplex you may be stuck with it

Hi,

please see the below output as you advice.This one will effect to network performance ?

 

GigabitEthernet9 is up, line protocol is up
Hardware is PQ3_TSEC, address is 00b7.716a.42db (bia 00b7.716a.42db)
Description: WAN NETWORK-ISP LINK
Internet address is 2.1.1.2/30
MTU 1500 bytes, BW 100000 Kbit/sec, DLY 100 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Half Duplex, 100Mbps, media type is RJ45
output flow-control is unsupported, input flow-control is unsupported
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: 1024
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 43000 bits/sec, 18 packets/sec
5 minute output rate 22000 bits/sec, 21 packets/sec
331312845 packets input, 112471702274 bytes, 0 no buffer
Received 14587 broadcasts (0 IP multicasts)
0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 0 multicast, 0 pause input
389785200 packets output, 102216124230 bytes, 10 underruns
6176889 output errors, 3229058 collisions, 1463886 interface resets
0 unknown protocol drops
0 babbles, 1820487 late collision, 4353333 deferred
6 lost carrier, 0 no carrier, 0 pause output
0 output buffer failures, 0 output buffers swapped out

 

 

Hi,

you told that Yes in theory it could stay up that long no problem but in the real world at some stage the ISP will introduce new hardware or do a software upgrade which would take down the circuit and that would cause the timers to reset ,whenever the BGP neighbour goes down ,the interface goes down or the circuit itself fails these would cause outages to the BGP relationship and reset the timers

 

That mean if bgp relationship outages ,it will effect connectivity ? how much outages should be acceptable ? Because my network bgp up time is aways change 1w ,3 w only. I never see 7 week or 8 wees. 

yes it will its your external link , so if it goes down and you dont have another connection it will cause an issue
1 week to 3weeks if thats constantly happening you need to investigate why , is it the ISP , do you have a bad circuit , is the physical link dropping or just the BGP neighbor , check exactly what the logs are showing you ,raise the issue with the circuit provider , try locate the problem is it your end or there's or somewhere in path, all depends on the downtime to but any hit can have knock on effects to applications that run at higher layers , they may not re converge so quick , so anything from a flap to a hard down is not good in reality for your network especially if critical services running

Hi ,

Thank for explains.you are correct some of my application got error.But I don't know how to ask ISP because I have less experience. I can verify my interface is always up.I can monitor the ISP physical link but i don't know how monitor the bgp peer issue.So let me know can I request to check their circuit and request log?

So let me know can I request to check their circuit and request log?
Yes you can , they would have provided a circuit ID you would need to call the ISP and request a circuit check through their management centre , some will provide logs there side some will not depending on the ISP but if you give them the times you had the outage they should be able to correlate to their end
You can also ask them to verify the stability of the circuit and explains you have had outages several times over a certain period
Your logs will also show if the BGP fails are coming from their end

good doc give you some things to look for when TSHOOT BGP
https://www.petri.com/bgp-troubleshooting-ccnp-route

Monitoring is a separate issue , you would either have to do it manually or else use software that can monitor the circuit (UIM, PRTG etc),and then email or alert you when its gone down,id set the logs to buffer 100000 so you can capture a lot of logs ,would also be a good idea to setup a syslog server to collect logs too

Hello MrBeginner,

late collisions are collisions that are detected after the first 64 bytes of the ethernet frame.

Usually, late collisions are caused by duplex mismatch : one side is full duplex and the other side is half duplex. Late collisions should increment on the half duplex side of the link.

It is strange that you see these errors after having changed the duplex settings on your side.

 

About BGP there is not an upper limit to a BGP session uptime.

As asked by Georg you should post the show interface <WAN> in order to understand the ratio of errors compared to valid packets. An error ratio less then 1% can be acceptable on half duplex link.

If the error ratio is greater then 2% there is some issue that needs to be solved starting to check the cable.

 

Hope to help

Giuseppe

 

 

Hi ,

How to calculate % .I am beginner in network.

Hello,

post the show interface <WAN>

and we will show you how to perform this calculation.

 

Hope to help

Giuseppe

 

Hi,

please see the below output as you advice.This one will effect to network performance ?

 

GigabitEthernet9 is up, line protocol is up
Hardware is PQ_TSEC, address is 00b7.716a.42db (bia 00b7.716a.42db)
Description: WAN NETWORK-ISP LINK
Internet address is 2.1.1.2/30
MTU 1500 bytes, BW 100000 Kbit/sec, DLY 100 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Half Duplex, 100Mbps, media type is RJ45
output flow-control is unsupported, input flow-control is unsupported
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: 1024
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 43000 bits/sec, 18 packets/sec
5 minute output rate 22000 bits/sec, 21 packets/sec
331312845 packets input, 112471702274 bytes, 0 no buffer
Received 14587 broadcasts (0 IP multicasts)
0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 0 multicast, 0 pause input
389785200 packets output, 102216124230 bytes, 10 underruns
6176889 output errors, 3229058 collisions, 1463886 interface resets
0 unknown protocol drops
0 babbles, 1820487 late collision, 4353333 deferred
6 lost carrier, 0 no carrier, 0 pause output
0 output buffer failures, 0 output buffers swapped out

 

 

Hello MrBeginner,

you have not cleared the counters.

We look at the output section

389785200 packets output

6176889 output errors,

 

the error ratio is calculated as : errored packets / (output packets + errored packets)=

= 6176889 / (6176889+389785200) = 0,01559969798017708710492230987295 = 1,559 %

 

However, to be noted there is an high number of interface resets :

>> 1463886 interface resets

This are not good also the late collisions if increasing should mean the other side is working in full duplex.

 

So check the LAN cable, change it with one that you know works well and try to use static settings

of

speed 100

duplex full

 

clear the counters on the interface with:

clear counters giga9

 

Note:

the use of speed 100 duplex full may solve or it can make the performance worse. Check with the ISP if their router supports 100 Mbps full duplex. If ISP router is able the issue should be on the LAN cable.

So changing the LAN cable is the first wise action to do.

 

Hope to help

Giuseppe

 

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: