sshd doesn't like this key: cat ~hansmi/.ssh/authorized_keys2 ssh-dss AAAB3NzaC1kc3MAAACBANuRyCB3NI4DEg5s9zWQID2N1zMUl/jKu8mjhhg03VbPUbIqhSy+5Eh0Bm/ccTdChBRvq6Y4pedFfkVhf5v2sAvhQfDB+q19kF2TLPUVvuKwClkr3Wk7+sSfhqlvxyEK04cwHkbM7fCopXFDhLcqaJuDsrVRM0FVfEFndsFtgPjPAAAAFQDUzCfxQxcZcPPQi6EGLNBhcSkewwAAAIEAhFtWzhgwHon0Ux8YhSdRy9mEDCT19Og6Eu/LkfqZEmhB38dJLzwDunNGRcAGMVvtm4280o8/bpqzCgJHh6/lq+pyBaP8gEdeFJRdZhuGcRc6ViW+hSxOc6fR/xNL+SSMxtDdSg925g6k79AyXGBnDA7r/REVpM3lLygDCEQhfqoAAACBAM0ZkrvhM6vq5lxXe9MrJA2bsM2laIBy michael@koala.lan when trying to login, the user gets this: $ ssh -l hansmi host -p 2222 Connection closed by 217.148.7.65 /var/log/auth.log says: Jan 15 12:59:33 aqua sshd[23772]: fatal: buffer_get: trying to get more bytes 129 than in buffer 24 password authentication works gcc-3.4.3, glibc-2.3.4.20041102-r0, 2.6.9-gentoo-r10 x86_64 ================================================================= System uname: 2.6.9-gentoo-r10 x86_64 AMD Opteron(tm) Processor 242 Gentoo Base System version 1.6.8 sys-devel/autoconf: 2.59-r6, 2.13 sys-devel/automake: 1.5, 1.4_p6, 1.9.4, 1.8.5-r2, 1.6.3, 1.7.9 sys-devel/binutils: 2.15.92.0.2-r2 sys-devel/libtool: 1.5.10-r2 virtual/os-headers: 2.6.8.1-r3 CFLAGS="-O3 -pipe -march=k8" CHOST="x86_64-pc-linux-gnu" CXXFLAGS="-O3 -pipe -march=k8" MAKEOPTS="-j3" it was built with support for the following: net-misc/openssh-3.9_p1-r1 tcpd -r1 is gentoo-specific this is an amd64-box please also have a look at the gentoo bug in the url
The key looks to be corrupt in some way (ssh-keygen can't parse it either). OpenSSH -current will more gracefully handle such broken keys and not terminate the connection when it encounters them. Could you check that the public key is not damaged? You can regenerate the public key from the private key using "ssh-keygen -yf /path/to/private.key"
seems like the key is correct, keygen built exactly the same again
Please attach the public key as a file to this bug.
Created attachment 768 [details] key here you are
The public key appears truncated: encoded it's about 140 bytes shorter than a 1024-bit DSA key. A couple of things to try: * does the OpenSSL library's self-test pass? ("make test" after building). * perhaps the private key is damaged. Does the same problem occur if you generate and use a new private key? * does the problem occur if you build OpenSSL and OpenSSH without the compiler optimization?
* does the OpenSSL library's self-test pass? ("make test" after building). yes * perhaps the private key is damaged. Does the same problem occur if you generate and use a new private key? that's possible, the same key has afaik also problems on a sparc/alpha. any other keys i'm using work, so i think it's the key that is truncated. thanks anyway
Change all RESOLVED bug to CLOSED with the exception of the ones fixed post-4.4.