Hi still same problem with the 7.0(3)I7(5) on a Nexus9000 C93180YC-FX   Log: 2019 May 9 15:39:24.389 N9000 %VPC-2-PEER_KEEP_ALIVE_RECV_FAIL: In domain 10, VPC peer keep-alive receive has failed 2019 May 9 15:39:24.389 N9000 %VPC-2-PEER_KEEP_ALIVE_SEND_FAIL: In domain 10, VPC peer keep-alive send has failed 2019 May 9 15:39:41.697 N9000 %AAA-6-AAA_ACCOUNTING_MESSAGE: update:10.241.21.32@pts/0:lebalch_t:Performing configuration copy. 2019 May 9 15:39:41.789 N9000 %SYSMGR-2-CFGWRITE_ABORTED_CONFELEMENT_RETRIES: Copy R S failed as config-failure retries are ongoing. Type "show nxapi retries" for checking the ongoing retries. 2019 May 9 15:39:41.791 N9000 %SYSMGR-3-CFGWRITE_SRVFAILED: Service "confelem" failed to store its configuration (error-id 0x00000079). 2019 May 9 15:39:42.692 N9000 %SYSMGR-2-CFGWRITE_ABORTED: Configuration copy aborted. 2019 May 9 15:39:43.001 N9000 %SYSMGR-3-CFGWRITE_FAILED: Configuration copy failed (error-id 0x401E0000).   Config VPC : N9000(config)# sh run vpc vpc domain 10 peer-switch role priority 10 system-priority 32667 peer-keepalive destination 10.246.19.115 source 10.246.19.114 udp-port 3200 vrf management interval 1000 timeout 5 precedence 6 hold-timeout 3 delay restore 200 no delay peer-link no dual-active exclude interface-vlan peer-gateway no layer3 peer-router layer3 peer-router syslog interval 3600 graceful consistency-check auto-recovery reload-delay 240 delay restore interface-vlan 10 delay restore orphan-port 0 no fast-convergence port-channel limit no shutdown ipv6 nd synchronize ip arp synchronize     VPC State : N9000(config)# sh vpc Legend: (*) - local vPC is down, forwarding vi a vPC peer-link vPC domain id : 10 Peer status : peer adjacency for med ok vPC keep-alive status : Suspended (Destina tion IP not reachable) Configuration consistency status : success Per-vlan consistency status : success Type-2 consistency status : success vPC role : secondary Number of vPCs configured : 0 Peer Gateway : Enabled Dual-active excluded VLANs : - Graceful Consistency Check : Enabled Auto-recovery status : Enabled, timer is off.(timeout = 240s) Delay-restore status : Timer is off.(time out = 200s) Delay-restore SVI status : Timer is off.(time out = 10s) Operational Layer3 Peer-router : Disabled vPC Peer-link status ------------------------------------------------------ --------------- id Port Status Active vlans -- ---- ------ ---------------------------------- --------------- 1 Po1 up 1,3900 ... View more

Thank you Paul   | Not so sure why you would want to if BGP RR capability is being advertised and is applicable to be used between peers   Ok I will elaborate.   Reminder :  sh ip bgp neighbors x.x.x.x routes <-- Routes received and learnt in BGP table. sh ip bgp neighbors x.x.x.x received-routes <-- All routes received even those filtered (not learnt in BGP table)   So why do we need to know ALL routes received (learnt AND not learnt) ? : We have a network AS, and around our network AS we have many other AS clients that are connecting to us via eBGP (we are transit between all of them) Then we need to know all the routes each AS client is trying to send to us. Based on this information we can then decide which route we going to accept or not in our BGP table. This kind of information is mandatory for us, because without the "received-routes" command we would be blind about our clients' AS routing tables. So of course we could ask our client to launch an "advertised-routes" on their side every time we need it, but you know process....reply time : 2-5 days, if you manage to get the client guy who understand what you need and have access to routers. So, possible but, no time, no way... -> With "received-routes" we get the same information instantly !!!   |  soft-reconfiguration is resource intensive to the router Not a problem in our network, it is private, this is not internet, in our corporation the number of route is around 10.000, no threat for our ASR1013...   Why we need route refresh ? : We want to be sure that our "received-routes" soft-database is up to date by "asking" our clients neighbor BGP router to resend their full advertisment.   Tommy ... View more

Thank you Paul for your response.   But still my question is : Why Cisco decided these 2 features to be exclusive ?   Why couldn't we have both feature at the same time ? Having a local database to store all  unedited routing informations (so that the show  "received-route" works) and Having a "Route Refresh" sent when we launch a clear bgp in ?   And because it is impossible, the next question is : If I disable the soft-reconfiguration how could I know le global list of "recieved-route" advertised by the neighbor peer ? This "received-route" information is very essential in our daily job because we have to know what a neighbor do advertise before we may accept and learn it.   Thanks. ... View more

Yes ! The solution have been to implement a 3rd Fabricpath domain connected to the first 2 (in triangle)   This 3rd fabricpath domain is made of 2 * 5548 in VPC, one on each site.   This 3rd fabripath domain is in RSTP   We connect all external LAN to this 3rd RSTP fabricpath domain. It is somewhat a DCI solution, not really supported by Cisco, but it works, and I think it is realiable enough for temporary migration purpose.   Tom ... View more

OK thank you, we will try it soon. I'll come back here for the results. ... View more

Hi ! I have the exact same problem, 2 customers A and B want to speak together via our network, but the 2 customers have their VPN with the same provider. (All routing is BGP so when the provider sees routes comming  from VPN A into VPN B he rejects the routes because in the AS-PATH the provider sees its own ASes.) How did you finally managed to "rewrite" the advertised as-path ? Thanks a lot Tom ... View more