# Downgrading for now since I can't reproduce this and my inbox suggests # nobody else is complaining about it just yet ... severity 344295 important tags 344295 moreinfo unreproducible thanks
On Wed, Dec 21, 2005 at 10:22:56AM -0600, Duane Meyer wrote: > This all seems to have started since the last update I ran on or > around the 19th of December 2005 which included a new 2.4.27 Kernel. > > Some time ago, I needed to set an environment variable to fix what > seemed to be a related problem. Nobody was able to connect via SSH to > the server. > > The fix seemed to relate to a double free of a memory chunk which > caused an updated glibc to abort the running program. So setting > MALLOC_CHECK_=0 would magically force glibc to ignore the error and > keep running. Maybe it's not related at all but the behavior looks > suspiciously famliar. As Steve says, we're probably going to need a backtrace to reproduce this. At a minimum, though, a copy of your /etc/ssh/sshd_config and information about any unusual configuration on your system would help. > $ sudo /usr/sbin/sshd -d [...] > debug1: list_hostkey_types: ssh-rsa,ssh-dss > Segmentation fault A segfault at this point indicates, I think, that it's within the first #ifdef GSSAPI branch in do_ssh2_kex() in sshd.c. I can't see anything much there that would segfault, though, certainly if you aren't using GSSAPI authentication ... 'sudo /usr/sbin/sshd -ddd' (more debugging) might help, although probably not a lot. Cheers, -- Colin Watson [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]