[Babel-users] Missing Neighbour

[Robert Lukan]

I found the problem. I dont know where is the real problem, but I know workaround for it.

I have script, that is working with serial devices, and making ppp connection over it, and assigning IP to that interface. And script assigned Ipv4 address when was pppX up(but not necessary connected with other client). And after examining log, I found this send: network is unreachable and after hour maybe two there was that error in memory allocation, but of course only when connection was unable to made.  

So when I changed my script to assign IPv4 when there was ALREADY assigned LL ipv6(so connection was made with other client, because pppd assign LL IPv6 when connection is made with other client), there was no problem with that. And it working very good, for now. Of course this error only occurred , when ppp was unable to make connection, so cable was disconnected.

I am not a programmer, but I think there is a problem where interface was going up and down, and it had assigned IPv4. So if interface don't have ipv4 assigned, babel is working OK, but connection with other side is not established. That mighty by just problem with interfaces like ppp. So there might be a memory leak ?!

My script assign ip with command: ifconfig pppx IP_ADRRESS netmask 255.255.255.255. And I have some special needs to have static assigned IPv4 address.
 
I can send slabinfo, if there is any need.
I know that is very special problem, but maybe sometimes will somebody found this useful ;)

If there is something unclear, just ask me. Or if you have any other suggestion.
Robert

-----Original Message-----
From: Juliusz Chroboczek [mailto:Juliusz.Chroboczek at pps.jussieu.fr]
Sent: Fri 8/1/2008 4:37 PM
To: Robert Lukan
Cc: babel-users at lists.alioth.debian.org
Subject: Re: [Babel-users] Missing Neighbour
 
> Should I send you a whole babel log file, I will compress it, maybe
> you will find something interesting. But as I said, there are no
> time stamps, just errors.

No, thanks.  As long as it's the kernel that is failing (ENOMEM from
IPV6_JOIN_GROUP), I cannot debug it from here.  I would assume
a kernel resource leak, either in the kernel or on Babel's side.

You may want to try with a different kernel release.  You may also
want to try to monitor /proc/slabinfo to see which resource is getting
exhausted.

Sorry, but there's really not much I can do on my side.

                                        Juliusz

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.alioth.debian.org/pipermail/babel-users/attachments/20080806/15a21864/attachment.htm