-
Notifications
You must be signed in to change notification settings - Fork 447
FAQ
Where can I ask for help ??
The best place to start is our google group (email [email protected]) and private Google + community (prevents spam and co)
But feel free to contact me privately if you want via mail and/or jabber at
> echo 'print ".".join(["exabgp@exa-networks","co","uk"])' | python
> echo 'print ".".join(["thomas","mangin@exa-networks","co","uk"])' | python
When can I ask for help ??
I normally answer as soon as I see the mails. My timezone is BST (United Kingdom).
Where is the documentation ?
Here and in the code .. really do you need more ??
If so please feel free to contribute some, I am a terrible author and my english despicable.
Is this code supported ?
Yes - Should you find any bugs, please report it, and I will fix them
I can not connect to ExaBGP
ExaBGP does not listen for incoming connection by default. You must specify the IP to listen to (default port is 179).
> env exabgp.tcp.bind="127.0.0.1" exabgp.tcp.port=1790 ../../../sbin/exabgp first.conf
Quagga is not happy with ExaBGP
ExaBGP will refuse BGP OPEN message with a router-id of 0.0.0.0 in accordance with RFC6286.
This is quagga's default router-id.
Quagga users need to setup their router-id to a non zero value when connecting to ExaBGP.
Does it work on Linux, MAC OS X, Windows ?
Linux: yes, it is our deployment environment
MACOS: yes, it is developed on MAC OSX 10.7
Windows: It should, but I have never tried. Let me know your experience.
How can I make exabgp reload its configuration ?
From 3.2 stable, send it a SIGUSR1, for example:
linux > kill -SIGUSR1 <pid>
BSD > kill -s SIGUSR1 <pid>
It forces ExaBGP to compare the previous and new configuration,
tear down and setup new peers, tear down sessions for un-configured ones,
and announce/withdraw the routes which changed.
Previous versions were using SIGHUP, it was changed to accommodate processes run within "screen".
"screen" sends SIGHUP to the program run when it terminates.
I keep seeing "could not connect to peer (if you use MD5, check your passwords): timed out" in the logs
OR your machine can not contact the router (firewall, routing issue, ...)
OR your MD5 passwords are not the same on both machines
OR you need to enable multi-hop BGP on your router
OR you are a FreeBSD user that has the incorrect password in /etc/ipsec.conf
I keep seeing "FreeBSD requires that you set your MD5 key via ipsec.conf." in the logs
You are attempting to configure a TCP MD5 Signature password from within the exabgp config file.
You must add 'md5 kernel;' to your configuration and follow the rest of the instructions on the
[examples] page.
I keep seeing "FreeBSD requires that you rebuild your kernel to enable TCP MD5 Signatures. Please review the documentation" in the logs
You are using FreeBSD, however, you have not recompiled your kernel based on the instructions.
I keep seeing "The client took over <number> seconds to send the OPEN, closing"
There is likely a problem with BGP Connection Collision Detection
Make sure that either ExaBPG router-id is BIGGER than your router configured router-id
or your router is in passive mode (does not try to establish outgoing connection).
For cisco 12.4 and later it is :
neighbor X.X.X.X transport connection-mode passive
For Juniper : set protocols bgp group group-name neighbor address passive
I keep seeing "This OS does not support TCP_MD5SIG, you can not use MD5 : Errno 42 Protocol not available" in the logs
You are most likely using a MAC and trying to use MD5 password
It does not work, if you need MD5 please use another OS
I keep seeing "Received Notification (3,10) from peer UPDATE message error: Invalid Network Field." in the logs
You are most likely trying to send IPv6 route(s) to a router which is not ready to accept them
You can check this by looking at the printed OPEN in the logs.
Or you are trying to send an IPv4 address on a session which only negotiated IPv6
This is caused as we send the route(s) even if the router did not announce the protocol
To fix it under junos type "set family inet6 unicast" under your bgp group or peer
I am trying to send some IPv6 routes to my router but the connection never stays up
You may need to enable family inet6 for this peer.
juniper:
set protocol bgp group mygroup neighbor IP family inet6 unicast
set protocol bgp group mygroup neighbor IP family inet unicast
Do not forget to explicitly enable IPv4 as well if you are planning to send both v6 and v4 routes
All seems fine but my route is not added to the routing table
You may need to allow for non-directly connected next-hop
juniper:
set protocol bgp group mygroup neighbor IP accept-remote-nexthop
Juniper does accept connection when I specify an HoldTime inferior to 20 seconds
Junipers will not accept a holdtime < 20 seconds ..
unless you enable a hidden command that lets you get it down to 6.
You can set it in the BGP protcols, group or neighbor level to set the configuration:
set minimum-hold-time 1
^
value 1 is not within range (6..65535) at '1'
I see I can specify the as-path (as-sequence) sent to the peer. Do you prepend the router ASN to peers ?
EBGP: we will add the router ASN automatically, to prevent BGP loops.
IBGP: it is used "as it".
Can I connect using a IPv6 connection ?
Yes, IPv6 is a first class citizen here :)
Can you maintain more than one BGP session ?
Yes, you can define more than one neighbor in the configuration file
Or have multiple instance running (like one instance per peer)
What is this "split" feature you can use on a route ?
exabgp allows you to write "route 10.0.0./16 next-hop 1.2.3.4 split /24"
This will then result in the announcement of the 256 x /24 routes which makes the /16
This is useful if you want to block large netmasks but to make sure your routes are used by the router
Can I save the program PID to use it with external scripts ?
Yes, just set the PID environment value with the location of the file
examples :
> env PID="/var/run/exabgp.pid" /opt/bin/bgpd /opt/etc/bgp/config.txt
> export PID="/var/run/exabgp.pid"
> /opt/bin/bgpd /opt/etc/bgp/config.txt
Are you planning to add more features ?
It has been a long time since I added a feature for my own use, so YES, I do try to help others !
Why did you implement Graceful Restart ?
My use case are machines announcing service IPs without lower MED/Local Prefence backup machines
ready to take the over in case of problems. The BGP IPs are then still routed during reboots and
in case of very high load for example it will not render the service IP unreacheable if some
keepalive are missed.
If no graceful-restart time is set, the waiting time is set to be the same as the hold-time.
As servers often take much longer than the default 180 seconds to reboots, we recommend that
you set the graceful-restart timer to a much higher value.
I would like to inject routes in the network but my server is behind a NATing firewall, will it work ?
NAT is not a good idea but: Yes, it will work.
Set your local-address as your machine LAN IP and the router-id as the firewall public IP.
As the code is single threaded, can a blocking network IO with one peer cause issues with another ?
We are using select to read data when available on the socket but we are not using a dedicated thread
for keepalive.
If it be a concern, run one instance of the program per peer.
High Disk IO (causing a long pause on configuration reload) may have the same effect,
that is why the default run script renice positively the program.
As you have no separate thread for keepalives, how many peers do you think one instance can manage comfortably?
it will really depend on the number of routes exchanged, and many other parameters,
but under normal use (even thousand of routes) you should never encounter an issue.
If unsure, just drop me a mail. I have heard of deployment with 160 peers
(low volume of routes update) .. and some others with one peers and insane number of update
(over 20/second constant).
Get Started
Misc
Self-Promotion