I configured a new hardware with RouterOS 2.9.41 and the same configuration that the other MK that works fine.
With the original router all works great, it have RouterOS 2.9.30.
When connect the new hardware with the same configuration, 4 of 6 bgp peer doesn´t established the connection:

Example:

name="PEER1" instance=default remote-address=200.XX.17.1 remote-as=11XXX tcp-md5-key="" nexthop-choice=default multihop=no route-reflect=no hold-time=3m ttl=1 in-filter="" out-filter=filter-peer1-out remote-id=0.0.0.0 remote-hold-time=0s used-hold-time=0s used-keepalive-time=0s state=active

07:23:52 route,bgp,info Failed to open TCP connection: No route to host
07:23:52 route,bgp,info RemoteAddr=200.XX.17.1
07:23:52 route,bgp,info RemotePort=179


The configuration is the same just as the other, what could it be ??.


Thanks a lot.

ttl=1 ... can you reach 200.xx.17.1 within 1 hop from that router ?

Sam

Yes Sam, in the other router (which are working now !) have the same configuration (with ttl=1).

07:22:57 route,bgp,info Failed to open TCP connection: No route to host
07:22:57 route,bgp,info RemoteAddr=200.xx.17.1
07:22:57 route,bgp,info RemotePort=179
07:23:17 route,bgp,debug,timer ConnectRetryTimer expired
07:23:17 route,bgp,debug,timer RemoteAddr=200.xx.17.1
07:23:17 route,bgp,debug Connecting
07:23:17 route,bgp,debug RemoteAddr=200.xx.17.1
07:23:17 route,bgp,debug RemotePort=179
07:23:17 route,bgp,debug,state Entering Connect state
07:23:17 route,bgp,debug,state RemoteAddr=200.xx.17.1
07:23:17 route,bgp,info Failed to open TCP connection: No route to host
07:23:17 route,bgp,info RemoteAddr=200.xx.17.1
07:23:17 route,bgp,info RemotePort=179
07:23:17 route,bgp,debug,state Entering Active state
07:23:17 route,bgp,debug,state RemoteAddr=200.xx.17.1
07:23:19 route,bgp,debug,timer KeepaliveTimer expired

anything changes if you set TTL more than 1?

07:22:57 route,bgp,info Failed to open TCP connection: No route to host
07:22:57 route,bgp,info RemoteAddr=200.xx.17.1
07:22:57 route,bgp,info RemotePort=179
07:23:17 route,bgp,debug,timer ConnectRetryTimer expired
07:23:17 route,bgp,debug,timer RemoteAddr=200.xx.17.1
07:23:17 route,bgp,debug Connecting
07:23:17 route,bgp,debug RemoteAddr=200.xx.17.1
07:23:17 route,bgp,debug RemotePort=179
07:23:17 route,bgp,debug,state Entering Connect state
07:23:17 route,bgp,debug,state RemoteAddr=200.xx.17.1
07:23:17 route,bgp,info Failed to open TCP connection: No route to host
07:23:17 route,bgp,info RemoteAddr=200.xx.17.1
07:23:17 route,bgp,info RemotePort=179
07:23:17 route,bgp,debug,state Entering Active state
07:23:17 route,bgp,debug,state RemoteAddr=200.xx.17.1
07:23:19 route,bgp,debug,timer KeepaliveTimer expired

07:23:17 route,bgp,info Failed to open TCP connection: No route to host -> What is that? Can you reach the 200.xx.17.1 with ICMP echo request (ping) ? Did you have routing information for that route? /ip route print ? I believe that your router doesn't know how to reach his BGP neighbor.

Kind Regards,
Danail Petrov

Danail, I can´t reach 200.xx.17.1 with ping.

Look the routing table:

# DST-ADDRESS PREF-SRC G GATEWAY DISTANCE INTERFACE
828 ADC 200.XX.17.0/24 200.XX.17.151 0 Peer1
829 Db 200.xX.17.0/24 r 200.XX.25.49 30 Peer2


And this is the routing table of my production router (it work fine):

# DST-ADDRESS PREF-SRC G GATEWAY DISTANCE INTERFACE


2081 ADC 200.XX.17.0/24 200.XX.17.151 0 Peer1
2082 Db 200.XX.17.0/24 r 200.XX.17.1 20 Peer1
2083 Db 200.XX.17.0/24 r 200.XX.17.2 20 Peer1

The Bgp against Peer2 work ok.

Thanks a lot

Danail, I can´t reach 200.xx.17.1 with ping.

Look the routing table:

# DST-ADDRESS PREF-SRC G GATEWAY DISTANCE INTERFACE
828 ADC 200.XX.17.0/24 200.XX.17.151 0 Peer1
829 Db 200.xX.17.0/24 r 200.XX.25.49 30 Peer2

Hi,
it looks like you have connection (Layer1,2) problems between two routers?That what i see, is one router with ip address 200.xx.17.1 (which is your production router) and the newest one 200.xx.17.151 which is in same broadcast domain as production router. So there is no routing decisions performed to connect them both , only layer2 ethernet connection must be fine. Can you see the ARP/IP association on both routers? Example: on your production router, try to /ip arp print to see what is in your ARP table. There must be an association with your hardware(MAC) address and the logical one (IP) like the output attached below:

# ADDRESS MAC-ADDRESS INTERFACE
0 D 192.168.11.10 00:20:ED:5E:06:25 LAN

Maybe i'm wrong , but it looks like you don't have layer2 connection between your routers. Maybe they are in different VLAN's , i don't know.But your problem is not in BGP protocol, i believe that you experiencing connection problems!


Kind Regards,
Danail Petrov

Danail, no:

200.XX.17.151 is my production router that work OK
200.XX.17.1 is the router of my provider (PEER1)

I´m trying to change a hardware for a new one and copy the configuration of the production router to this new one (and with which have the problem).
Do you understand me ?.
Yes, I supposed that have layer 1,2 problem, but I can´t troubleshoot the problem !.


Thanks

Janisk, I didn´t change the ttl, I´m going to try this tomorrow, what value must have ?. Remember that my production router have the ttl=1 and is a copy of them.

Janisk, I didn´t change the ttl, I´m going to try this tomorrow, what value must have ?. Remember that my production router have the ttl=1 and is a copy of them.

The TTL value my friend, is an IP option. You don't have IP problems. IP operates at layer3 of OSI model,until you have layer1/2 problems, there is nothing you can do with IP. Check your physical connection (wires and switches)

BR

I need to update the PC configuration.

i got working right now a PC with 4 NICs.
i wanted to upgrade this very old equipment with a new one, with the addition of 1 more NIC.

i have tested several motherboards earlier with no success:

the biggest problem is to find some MB with 5 PCI slots.
when finding such MB, some of the NICs were not detected by RouterOS.

i finally found an asus MB.

all NICs seemed to be working, but i got some strange beheavior, descripted above.

could it be hardware related?
is there a problem with number of the NICs and the PCI bus/IRQ´s etc ?


Thanks a lot

I need to update the PC configuration.

i got working right now a PC with 4 NICs.
i wanted to upgrade this very old equipment with a new one, with the addition of 1 more NIC.

i have tested several motherboards earlier with no success:

the biggest problem is to find some MB with 5 PCI slots.
when finding such MB, some of the NICs were not detected by RouterOS.

i finally found an asus MB.

all NICs seemed to be working, but i got some strange beheavior, descripted above.

could it be hardware related?
is there a problem with number of the NICs and the PCI bus/IRQ´s etc ?


Thanks a lot

I have one more suspicion, are you sure that logical addresses (IP) are configured on the right ethernet adapters ? I mean , if you have 4 ethernet cards , you may easy getting confused about which one is configured and which one is actually attached to network?

To your other questions , i believe that there's not IRQ issues , because they can be shared when same kind of hardware is used (ethernet cards)

BR,
Danail Petrov

I´m using 4 NIC´s 3COM 3c905C-TX/TX-M [Tornado] and connect him directly to peers, the link is OK, what could it be ??.


Thanks

When monitoring links status I saw "link-ok" and then "no-link" (flaps), I supposed that the problem is the 3COM cards, could it be ??, could be the 2.9.41 version ?.

Please help me.


Thanks a lot.

filter all incoming routes and see if that makes it more stable. maybe you have a looping route or something misconfigured. I've seen specific routes cause problems (flaps, etc).

Sam

Finally the problem was the 3COM cards, now use Realtek ..., I can´t understand what´s happend.



Thanks