03-27-2014 03:13 AM - edited 03-04-2019 10:39 PM
Hi Team,
I have Cisco 3825 router and configured a Gig0/0 for the internet link. Recently we have observed that some of the sites like facebook/sbi/supportforums.cisco.com not opening. Whereas i am able to open all other sites like cisco.com/twitter.com
Can someone please help me on this
interface GigabitEthernet0/1
mtu 1500
ip address x.x.x.x. 255.255.255.252
ip virtual-reassembly
no ip route-cache cef
no ip route-cache
ip tcp adjust-mss 1450
load-interval 30
duplex auto
speed auto
media-type rj45
hold-queue 524 in
hold-queue 524 out
|
router bgp xxxx
bgp log-neighbor-changes
neighbor x.x.x.x remote-as x.x.x.x
neighbor update-source GigabitEthernet0/1
maximum-paths 2
Regards
MAC
03-27-2014 04:55 AM
It may be an MTU issue... Have you tried changing the MTU to reflect the one set by your ISP?
03-27-2014 05:09 AM
Hi Nish,
Yes i have tried changing the MTU and also the tcp MSS value but no luck
Will changing the value in Weight will do any difference?
neighbor x.x.x.x weight 550
Regards
MAC
03-27-2014 05:32 AM
How many BGP peers do you have? 1?
Weight is locally significant and also check the hold-queue, as "Cisco recommends that the queue size not exceed ten packets on asynchronous interfaces"
I would do a packet capture and see whats going on...
03-27-2014 05:34 AM
Hi Nish,
We have only one BGP Peers, some few days back we were having two but i have removed the unwanted configuration.
As of now MTU size is 1500, i tried changing it but no luck
Regards
Asif C Y
03-27-2014 05:49 AM
Can you please explain why the following is there...? Have a packet capture and see whats going on...
router bgp xxxx
bgp log-neighbor-changes
neighbor x.x.x.x remote-as x.x.x.x
neighbor update-source GigabitEthernet0/1
maximum-paths 2
03-27-2014 06:12 AM
Hi Nish,
The config you are referring has been removed for now and tested. But no proper results, i am trying to run debug commands and take results
Regards
Asif C Y
03-27-2014 06:38 AM
Can you try TCP MSS to 1452 and double check the MTU, I really think one of them is causing it...
03-27-2014 07:08 AM
I do not know that it would cause this problem but I question why this is configured on the interface
no ip route-cache cef
Using CEF usually improves performance and when you turn it off you should result in all packets being process switched which does impact performance.
I also wonder why this is on the interface
ip virtual-reassembly
and it makes me think that fragmentation and reassembly may be the issue. As a test I suggest that you use the ip tcp adjust-mss to 1300 and see if the behavior changes.
HTH
Rick
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide