Bug #14500
closed"Error reading SSH protocol banner" in rados-infernalis-distro-basic-openstack
0%
Description
I see these errors in many runs, but jobs seem to be recovering after it and running fine, so maybe benign
Run: http://pulpito.ovh.sepia.ceph.com:8081/teuthology-2016-01-23_02:00:01-rados-infernalis-distro-basic-openstack/
Jobs: all
Logs: http://teuthology.ovh.sepia.ceph.com/teuthology/teuthology-2016-01-23_02:00:01-rados-infernalis-distro-basic-openstack/8006/teuthology.log
2016-01-23T02:57:22.835 INFO:teuthology.misc:Re-opening connections... 2016-01-23T02:57:22.835 INFO:teuthology.misc:trying to connect to ubuntu@target085224.ovh.sepia.ceph.com 2016-01-23T02:57:22.837 DEBUG:teuthology.orchestra.connection:{'username': 'ubuntu', 'hostname': 'target085224.ovh.sepia.ceph.com', 'key_filename': '/home/teuthworker/.ssh/id_rsa', 'timeout': 60} 2016-01-23T02:57:23.402 ERROR:paramiko.transport:Exception: Error reading SSH protocol banner[Errno 104] Connection reset by peer 2016-01-23T02:57:23.403 ERROR:paramiko.transport:Traceback (most recent call last): 2016-01-23T02:57:23.404 ERROR:paramiko.transport: File "/home/teuthworker/src/teuthology_master/virtualenv/local/lib/python2.7/site-packages/paramiko/transport.py", line 1516, in run 2016-01-23T02:57:23.404 ERROR:paramiko.transport: self._check_banner() 2016-01-23T02:57:23.404 ERROR:paramiko.transport: File "/home/teuthworker/src/teuthology_master/virtualenv/local/lib/python2.7/site-packages/paramiko/transport.py", line 1633, in _check_banner 2016-01-23T02:57:23.404 ERROR:paramiko.transport: raise SSHException('Error reading SSH protocol banner' + str(x)) 2016-01-23T02:57:23.405 ERROR:paramiko.transport:SSHException: Error reading SSH protocol banner[Errno 104] Connection reset by peer 2016-01-23T02:57:23.405 ERROR:paramiko.transport: 2016-01-23T02:57:23.426 DEBUG:teuthology.orchestra.remote:Error reading SSH protocol banner[Errno 104] Connection reset by peer
Updated by Yuri Weinstein about 8 years ago
Issues are still here, but seem not harmful
http://qa-proxy.ceph.com/teuthology/teuthology-2016-04-05_22:00:02-rados-jewel-distro-basic-smithi/110414/teuthology.log
Updated by Zack Cerza about 8 years ago
- Status changed from New to Closed
Either it's network noise, or something that might have been fixed by the paramiko upgrade. Let's reopen if we continue to see this.
Updated by Brad Hubbard almost 5 years ago
A note to anyone seeing this currently. "Error reading SSH protocol banner" is kind of a catch-all error for paramiko so the causes can be various. In my case the actual, underlying error was the following.
Traceback (most recent call last): File "test_connect.py", line 23, in <module> ssh.connect(**connect_args) File "/home/bhubbard/src/teuthology/virtualenv/local/lib/python2.7/site-packages/paramiko/client.py", line 414, in connect raise BadHostKeyException(hostname, server_key, our_key) paramiko.ssh_exception.BadHostKeyException: ('miraX.XXX.XXX.ceph.com', <paramiko.ecdsakey.ECDSAKey object at 0x7fb534521890>, <paramiko.ecdsakey.ECDSAKey object at 0x7fb534ec8a10>)
This was due to conflicting entries for the machines in my known_hosts file and once they were removed the operation succeeded. After my analysis of the issue the Sepia wiki [1] was updated to recommend use of the "UserKnownHostsFile /dev/null" directive in ssh/config.
[1] https://wiki.sepia.ceph.com/doku.php?id=testnodeaccess#sshexceptionerror_reading_ssh_protocol_banner