Re: node 0.3.1 connect_to: connect: No buffer space available

From: Paul Lewis (elqq.snpowbcn@icmc.usp.br)
Date: Sun May 04 2003 - 12:48:09 EEST

  • Next message: John: "Age and the art of making a EXT3 File System Kernel"

    Hi Riley

    In message <crnh.kcmvegop@rele.tunk.net>, Riley
    Williams <dtl.zrcw@mpsaz.org> writes
    >Hi Paul.
    >
    > >> Anyone got pointers to this error.
    > >>
    > >> Internet search has not come up with an answer.
    > >> Do I have control over the buffer referred to?
    > >>
    > >> Occurred since I added encap routes to my server's gateway
    > >>
    > >> Q> T 44.131.10.55
    > >> Q> GB7CR:CRNODE CRCHAT:GB7CIP-5 connect_to: connect:
    > >> Q> No buffer space available
    >
    > > Still tracking this down, not a node error after searching around
    > > the Internet for clues. Seems to be a general error message.
    > >
    > > Eventually located the following in the kernel log in the early
    > > hours.....Lots of these messages
    > >
    > > Now to find out how many 'neighbour' are allowed. may be related to
    > > the number of active interfaces I have operational on my system
    > >
    > > May 3 20:29:52 gb7cip kernel: neighbour table overflow
    > > May 3 20:29:57 gb7cip kernel: neighbour table overflow
    > > May 3 20:30:02 gb7cip kernel: neighbour table overflow
    > > May 3 20:30:07 gb7cip kernel: neighbour table overflow
    > > May 3 20:30:12 gb7cip kernel: neighbour table overflow
    > > May 3 20:30:17 gb7cip kernel: neighbour table overflow
    > > May 3 20:30:22 gb7cip kernel: neighbour table overflow
    >
    >This sounds like a fairly well known problem connected with ARP. The
    >Linux kernel has a smallish limit on the number of ARP entries it can
    >handle, and one needs to run arpd (the ARP daemon) if one expects to
    >need more than that.
    >
    >ARP as referred to above has to do with TCP/IP - specifically with the
    >IP part thereof - and is primarily connected with Ethernet interfaces.
    >However, the problem it solves is equally applicable to AX.25 in its
    >NETROM and TCP/IP-over-AX.25 modes, so could be related.
    >
    >Also, a message every five seconds? Looks like it's been rate-limited.
    >That tends to be done for messages where an error tends to reoccur once
    >it first occurs, so it's obviously something that's been foreseen...
    >
    >Best wishes from Riley G7GOD / KB8PPG

    Thanks for the tip on the arpd demon, saw reference to that on an Inet
    search. Though it was a message a few years old.

    I agree it's pointing to the Arp table (now known as neighbour - took me
    awhile to track (remember that)

    I will try that, sounds like the answer I am looking for.

    Found this document that gave me some more upto date background

    http://www.fokus.gmd.de/linux/HOWTO/pdf/Adv-Routing-HOWTO.pdf
    Paul g4apl 

    -- 
neqwsobv.acephpexm@kerailya.tunkki.fi
-
To unsubscribe from this list: send the line "unsubscribe linux-hams" in
the body of a message to terhi.victor@logonet.com
More majordomo info at  http://vger.kernel.org/majordomo-info.html

    This archive was generated by hypermail 2b30 : Sun May 04 2003 - 12:57:54 EEST