I had the same problem myself - upping to 2.0.13 corrected this. The
problem I had, more specifically, was that a sshd wouldn't die, even if a
connection was never successfully established.
With 2.0.12, this also happened because of something in a remote user's
.ssh2 directory. That is, when he came from server B to server A, the
session on server A would hang when he exits. And he solved this by
delketing his .ssh2 directory, so we never did track it down.
--
Gregor Mosheh
[EMAIL PROTECTED]
On-Site Systems Admin, Humboldt Internet
707.825.4638
On Tue, 28 Sep 1999, [ISO-8859-1] P�r Lanvin wrote:
>
> Hi,
>
> Versions:
> OS: Solaris 2.6
> HW: SPARC and PC (Compaq)
> SSH server: 2.1.12 (SPARC)
> SSH client: 2.1.12 (SPARC) and 2.0.8, build 7 (Windows95)
>
>
> This may seem a little complicated but it is not..
>
> Problem:
> The SSH daemon rather often spawns new SSH daemons which go wild thus
> consuming a lot of CPU.
>
> Env:
> The machine that runs the daemon is a SPARC w/ 2.6,.
> Clients who connects to it run Solaris on SPARC.
> The PC clients are on a separate Novell network w/ w95 on Compaq.
> Connectin w/ ReflectionX (Yes, I know ReflextionX totally lacks X
> authorization).
>
> I am pretty sure that it is not the wirus95 clients which are the
> cause of this problem but the clients from the SPARCs.
>
> Have anyone else seen this problem ? is there a fix ?
> I upgraded to .13 today, dunno if it helps.
>
> My only theory so far is that it might involve bash 1.0.16.
> Will try bash 2.x if the SSH upgrade does not do any good.
>
>
>
>
> rgds
>
> /PL
>
>
> P Lanvin
> 46&2 NSA
> Network Security and Administration
>
>
