New user, initial question
mrenzmann at compulan.de
Fri Dec 29 13:27:42 CET 2000
On Fri, 29 Dec 2000 11:39:33 +0100
Florian Lohoff <flo at rfc822.org> wrote:
> I didnt like zebra because of the multiple deamon approach. This might
> be easier to track down bugs and better do debuggin but when i say
> "Dynamic Routing" i dont want to start another 10 daemons each listing
> on its own port for configuration. My experience with zebra
> is 1-2 years old and this might have changed since then but i like bird.
I got some simple problems, but did not have enough time to fix them by myself
and the zebra team seems to be busy with other stuff (at least I did not
receive any answer to my questions). The most annoying problem was that
sending a "SIGHUP" to ripd (the RIPv2 portion of zebra) caused it to reopen
the port RIP listens to. As this port never has been closed before, an
unprivileged port gets used instead. And that is crap. But I have to send a
SIGHUP from time to time as soon as the configuration has been changed. I
worked around this problem by shutting down ripd first and restarted it then.
But this also is crap.
Another issue was that the documentation has not been clear on some things I
needed, and it was quite a bit complicated for me to find the hints I needed.
My first impression of the Bird docu is much(!) better. Quite understandable,
> Go for it - I have seen memory leaks in 1.0.0 and on - They definitly
> Beside the tiny, not so clear memory leak, i havent had time to debug,
> the RipV2 part is ROCKsolid.
> Ah - just short idea - birdc and do a
> "dump resources" and immediatly the bird catches up on the memory
> mgr1:~# ps auxww | grep bird
> root 5231 0.0 11.6 15804 14804 ? S Sep06 54:49
I will try to reproduce it on my local machines as soon as bird runs on them.
More information about the Bird-users