--JcvBIhDvR6w3jUPA
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Jul 10, 2006 at 09:43:13PM +0200, Jan Danielsson wrote:
> Hello all,
>=20
>    Sometimes, when I really push my network interface to its limit, I
> get the message:
>=20
> rtk0: unable to allocate Tx cluster

rtw or rtk, or Rx?  rtw(4) can generate a Tx cluster message, rtk can
have a Rx cluster message.  (assuming 3.0 sources)

Anyway if this is the rtk Rx cluster error, it's a kernel memory/mbuf
management issue, the GATEWAY option bumps NMBCLUSTERS, perhaps this
is necessary in this case. (to pull a suggestion out of thick air)

>=20
>    ...on the console.
>=20
>    I thought I would do some good for mankind, so I installed tor -- as
> a server. Pretty soon I found out that being a server in the tor network
> implies *very* *much* traffic through your node.
>=20
>    This *really* pushed my network interface to its limits. When I came
> home a few minutes ago, I found that the console was filled with the
> message above, and it was spewing out new ones all the time. The ONLY
> thing I can do with the system is to switch terminals
> (Ctrl+Alt+F{1,2,3,4}), but I can not type any user name to login. It is
> not reachable via ssh.
>=20
>    So, I need to reset it. :-(
>=20
>    (Pushing the power button got me a bunch of "syncing disks ... 4 4 4
> 4 4 4 4 4 4 4 4 4 4 ..." and some message I didn't have time to read.)
>=20
>    Is there some known problem with the rtk driver, or is it the chip on
> my network card?

This chip is generally regarded as inferior, but FWIW I have never had any
problems with mine.

	Jonathan Kollasch

--JcvBIhDvR6w3jUPA
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (NetBSD)

iD8DBQFEsxtuOjx1ye3hmokRAmU+AJ9ntAQh46PSXQT2gcdIoSI99zotwwCfQzvd
42bcUHYuQylLH4w3Gz7oLjM=
=b5Ck
-----END PGP SIGNATURE-----

--JcvBIhDvR6w3jUPA--