ssh-select

SSH-select is an auto-completion wrapper for SSH. The completion engine reads hosts from various known locations for SSH known_hosts files, but can be instructed to search other locations as well.

SSH-select is a one-off wrapper, i.e. once a host has been selected, its process is replaced with that of an SSH client using the exec system call. There are no daemons or other lingering processes involved.

examples

ssh-select

without any arguments, known locations will be searched for host inputs. the resulting suggestions vary between execution environments.

ssh-select --no-search --known-hosts /tmp/ssh_known_hosts

populates the suggestion database with entries from the provided known_hosts file. if the file does not exist, no autocompletion is possible as other sources have been disabled.

ssh-select --ssh /opt/bin/rsh

passes the collected arguments to the provided binary instead of the regular ssh binary.

ssh-select --no-search --hosts /etc/hosts

disables the automated import from well known host sources, but adds the /etc/hosts file back into the selection.

ssh-select --zone /var/named/example.com.zone

reads hostnames from the provided zone file.

ssh-select --zone '/var/lib/bind/**/db.*'

use hostnames from each zone file the glob pattern resolves to.

NOTE: when using globs, wrap the argument in quotes to prevent the shell from expanding it, which would cause only a single file to be used as input and the rest as ssh arguments

ssh-select --named-dump /var/lib/bind/named_dump.db

reads hostnames from the Bind zone dump file. The file is usually generated using rndc (e.g. rndc dumpdb -zones)

ssh-select -i /tmp/id_rsa

arguments not supported by ssh-select will be forwarded to ssh as-is.

configuration

SSH-select can change its behaviour base on commandline arguments and environment variables. commandline arguments which are not understood/supported are passed on to the spawned SSH client unchanged.

commandline arguments

commandline arguments take precedence over environment variables:

• --ssh

  path to the SSH client to invoke once a host has been selected. if not specified, ssh is searched for in the users PATH.

• --known-hosts

  path to a known_hosts file. can be specified multiple times. duplicate hosts are filtered. supports glob patterns.

• --hosts

  path to a hosts file. can be specified multiple times. duplicate hosts are filtered. supports glob patterns.

• --zone

  path to a domain name zone file. can be specified multiple times. duplicate hosts are filtered. supports glob patterns.

  currently only the following resource records are processed:

  PTR, A, AAAA, CNAME, MX, NS

• --named-dump path to a Bind zone dump file. can be specified multiple times. duplicate hosts are filtered. supports glob patterns.

  currently only the following resource records are processed:

  PTR, A, AAAA, CNAME, MX, NS

• --no-search

  omit adding hosts from well-known file locations.

• --no-op

  do not actually execute ssh

• --quiet

  omit the search summary upon start.

environment variables

environment variables related to the application are prefixed with SSH_SELECT_:

• SSH_SELECT_NO_SEARCH

  if defined, has the same effect as --no-search.

• SSH_SELECT_NO_OP

  if defined, has the same effect as --no-op.

• SSH_SELECT_QUIET

  if defined, has the same effect as --quiet.

• SSH_SELECT_SSH_BINARY

  same as --ssh

• SSH_SELECT_KNOWN_HOSTS_FILE_$

  environment variables with this prefix have the same effect as --known-hosts; the suffix has no effect on the result.

• SSH_SELECT_HOSTS_FILE_$

  environment variables with this prefix have the same effect as --hosts; the suffix has no effect on the result.

• SSH_SELECT_ZONE_FILE_$

  environment variables with this prefix have the same effect as --zone; the suffix has no effect on the result.

• SSH_SELECT_NAMED_DUMP_$

  environment variables with this prefix have the same effect as --named-dump; the suffix has no effect on the result.