- (djm) [contrib/ssh-copy-id contrib/ssh-copy-id.1] Updated to Phil

Hands' greatly revised version.
This commit is contained in:
Damien Miller 2013-03-22 10:17:36 +11:00
parent 63b4bcd04e
commit 83efe7c861
3 changed files with 470 additions and 116 deletions

View File

@ -1,3 +1,7 @@
20120322
- (djm) [contrib/ssh-copy-id contrib/ssh-copy-id.1] Updated to Phil
Hands' greatly revised version.
20120318 20120318
- (djm) [configure.ac log.c scp.c sshconnect2.c openbsd-compat/vis.c] - (djm) [configure.ac log.c scp.c sshconnect2.c openbsd-compat/vis.c]
[openbsd-compat/vis.h] FreeBSD's strnvis isn't compatible with OpenBSD's [openbsd-compat/vis.h] FreeBSD's strnvis isn't compatible with OpenBSD's

View File

@ -1,54 +1,293 @@
#!/bin/sh #!/bin/sh
# Shell script to install your public key on a remote machine # Copyright (c) 1999-2013 Philip Hands <phil@hands.com>
# Takes the remote machine name as an argument. # 2013 Martin Kletzander <mkletzan@redhat.com>
# Obviously, the remote machine must accept password authentication, # 2010 Adeodato =?iso-8859-1?Q?Sim=F3?= <asp16@alu.ua.es>
# or one of the other keys in your ssh-agent, for this to work. # 2010 Eric Moret <eric.moret@gmail.com>
# 2009 Xr <xr@i-jeuxvideo.com>
# 2007 Justin Pryzby <justinpryzby@users.sourceforge.net>
# 2004 Reini Urban <rurban@x-ray.at>
# 2003 Colin Watson <cjwatson@debian.org>
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions
# are met:
# 1. Redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer.
# 2. Redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution.
#
# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
# OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
# IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
# THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
ID_FILE="${HOME}/.ssh/id_rsa.pub" # Shell script to install your public key(s) on a remote machine
# See the ssh-copy-id(1) man page for details
if [ "-i" = "$1" ]; then # check that we have something mildly sane as our shell, or try to find something better
shift if false ^ printf "%s: WARNING: ancient shell, hunting for a more modern one... " "$0"
# check if we have 2 parameters left, if so the first is the new ID file then
if [ -n "$2" ]; then SANE_SH=${SANE_SH:-/usr/bin/ksh}
if expr "$1" : ".*\.pub" > /dev/null ; then if printf 'true ^ false\n' | "$SANE_SH"
ID_FILE="$1" then
printf "'%s' seems viable.\n" "$SANE_SH"
exec "$SANE_SH" "$0" "$@"
else else
ID_FILE="$1.pub" cat <<-EOF
oh dear.
If you have a more recent shell available, that supports \$(...) etc.
please try setting the environment variable SANE_SH to the path of that
shell, and then retry running this script. If that works, please report
a bug describing your setup, and the shell you used to make it work.
EOF
printf "%s: ERROR: Less dimwitted shell required.\n" "$0"
exit 1
fi fi
shift # and this should leave $1 as the target name fi
DEFAULT_PUB_ID_FILE=$(ls -t ${HOME}/.ssh/id*.pub 2>/dev/null | grep -v -- '-cert.pub$' | head -n 1)
usage () {
printf 'Usage: %s [-h|-?|-n] [-i [identity_file]] [-p port] [[-o <ssh -o options>] ...] [user@]hostname\n' "$0" >&2
exit 1
}
# escape any single quotes in an argument
quote() {
printf "%s\n" "$1" | sed -e "s/'/'\\\\''/g"
}
use_id_file() {
local L_ID_FILE="$1"
if expr "$L_ID_FILE" : ".*\.pub$" >/dev/null ; then
PUB_ID_FILE="$L_ID_FILE"
else
PUB_ID_FILE="$L_ID_FILE.pub"
fi fi
PRIV_ID_FILE=$(dirname "$PUB_ID_FILE")/$(basename "$PUB_ID_FILE" .pub)
# check that the files are readable
for f in $PUB_ID_FILE $PRIV_ID_FILE ; do
ErrMSG=$( { : < $f ; } 2>&1 ) || {
printf "\n%s: ERROR: failed to open ID file '%s': %s\n\n" "$0" "$f" "$(printf "%s\n" "$ErrMSG" | sed -e 's/.*: *//')"
exit 1
}
done
GET_ID="cat \"$PUB_ID_FILE\""
}
if [ -n "$SSH_AUTH_SOCK" ] && ssh-add -L >/dev/null 2>&1 ; then
GET_ID="ssh-add -L"
fi
while test "$#" -gt 0
do
[ "${SEEN_OPT_I}" ] && expr "$1" : "[-]i" >/dev/null && {
printf "\n%s: ERROR: -i option must not be specified more than once\n\n" "$0"
usage
}
OPT= OPTARG=
# implement something like getopt to avoid Solaris pain
case "$1" in
-i?*|-o?*|-p?*)
OPT="$(printf -- "$1"|cut -c1-2)"
OPTARG="$(printf -- "$1"|cut -c3-)"
shift
;;
-o|-p)
OPT="$1"
OPTARG="$2"
shift 2
;;
-i)
OPT="$1"
test "$#" -le 2 || expr "$2" : "[-]" >/dev/null || {
OPTARG="$2"
shift
}
shift
;;
-n|-h|-\?)
OPT="$1"
OPTARG=
shift
;;
--)
shift
while test "$#" -gt 0
do
SAVEARGS="${SAVEARGS:+$SAVEARGS }'$(quote "$1")'"
shift
done
break
;;
-*)
printf "\n%s: ERROR: invalid option (%s)\n\n" "$0" "$1"
usage
;;
*)
SAVEARGS="${SAVEARGS:+$SAVEARGS }'$(quote "$1")'"
shift
continue
;;
esac
case "$OPT" in
-i)
SEEN_OPT_I="yes"
use_id_file "${OPTARG:-$DEFAULT_PUB_ID_FILE}"
;;
-o|-p)
SSH_OPTS="${SSH_OPTS:+$SSH_OPTS }$OPT '$(quote "$OPTARG")'"
;;
-n)
DRY_RUN=1
;;
-h|-\?)
usage
;;
esac
done
eval set -- "$SAVEARGS"
if [ $# != 1 ] ; then
printf '%s: ERROR: Too many arguments. Expecting a target hostname, got: %s\n\n' "$0" "$SAVEARGS" >&2
usage
fi
# drop trailing colon
USER_HOST=$(printf "%s\n" "$1" | sed 's/:$//')
# tack the hostname onto SSH_OPTS
SSH_OPTS="${SSH_OPTS:+$SSH_OPTS }'$(quote "$USER_HOST")'"
# and populate "$@" for later use (only way to get proper quoting of options)
eval set -- "$SSH_OPTS"
if [ -z "$(eval $GET_ID)" ] && [ -r "${PUB_ID_FILE:=$DEFAULT_PUB_ID_FILE}" ] ; then
use_id_file "$PUB_ID_FILE"
fi
if [ -z "$(eval $GET_ID)" ] ; then
printf '%s: ERROR: No identities found\n' "$0" >&2
exit 1
fi
# populate_new_ids() uses several global variables ($USER_HOST, $SSH_OPTS ...)
# and has the side effect of setting $NEW_IDS
populate_new_ids() {
local L_SUCCESS="$1"
# repopulate "$@" inside this function
eval set -- "$SSH_OPTS"
umask 0177
local L_TMP_ID_FILE=$(mktemp ~/.ssh/ssh-copy-id_id.XXXXXXXXXX)
trap "rm -f $L_TMP_ID_FILE*" EXIT TERM INT QUIT
printf '%s: INFO: attempting to log in with the new key(s), to filter out any that are already installed\n' "$0" >&2
NEW_IDS=$(
eval $GET_ID | {
while read ID ; do
printf '%s\n' "$ID" > $L_TMP_ID_FILE
# the next line assumes $PRIV_ID_FILE only set if using a single id file - this
# assumption will break if we implement the possibility of multiple -i options.
# The point being that if file based, ssh needs the private key, which it cannot
# find if only given the contents of the .pub file in an unrelated tmpfile
ssh -i "${PRIV_ID_FILE:-$L_TMP_ID_FILE}" \
-o PreferredAuthentications=publickey \
-o IdentitiesOnly=yes "$@" exit 2>$L_TMP_ID_FILE.stderr </dev/null
if [ "$?" = "$L_SUCCESS" ] ; then
: > $L_TMP_ID_FILE
else
grep 'Permission denied' $L_TMP_ID_FILE.stderr >/dev/null || {
sed -e 's/^/ERROR: /' <$L_TMP_ID_FILE.stderr >$L_TMP_ID_FILE
cat >/dev/null #consume the other keys, causing loop to end
}
fi
cat $L_TMP_ID_FILE
done
}
)
rm -f $L_TMP_ID_FILE* && trap - EXIT TERM INT QUIT
if expr "$NEW_IDS" : "^ERROR: " >/dev/null ; then
printf '\n%s: %s\n\n' "$0" "$NEW_IDS" >&2
exit 1
fi
if [ -z "$NEW_IDS" ] ; then
printf '\n%s: WARNING: All keys were skipped because they already exist on the remote system.\n\n' "$0" >&2
exit 0
fi
printf '%s: INFO: %d key(s) remain to be installed -- if you are prompted now it is to install the new keys\n' "$0" "$(printf '%s\n' "$NEW_IDS" | wc -l)" >&2
}
REMOTE_VERSION=$(ssh -v -o PreferredAuthentications=',' "$@" 2>&1 |
sed -ne 's/.*remote software version //p')
case "$REMOTE_VERSION" in
NetScreen*)
populate_new_ids 1
for KEY in $(printf "%s" "$NEW_IDS" | cut -d' ' -f2) ; do
KEY_NO=$(($KEY_NO + 1))
printf "%s\n" "$KEY" | grep ssh-dss >/dev/null || {
printf '%s: WARNING: Non-dsa key (#%d) skipped (NetScreen only supports DSA keys)\n' "$0" "$KEY_NO" >&2
continue
}
[ "$DRY_RUN" ] || printf 'set ssh pka-dsa key %s\nsave\nexit\n' "$KEY" | ssh -T "$@" >/dev/null 2>&1
if [ $? = 255 ] ; then
printf '%s: ERROR: installation of key #%d failed (please report a bug describing what caused this, so that we can make this message useful)\n' "$0" "$KEY_NO" >&2
else
ADDED=$(($ADDED + 1))
fi
done
if [ -z "$ADDED" ] ; then
exit 1
fi
;;
*)
# Assuming that the remote host treats ~/.ssh/authorized_keys as one might expect
populate_new_ids 0
[ "$DRY_RUN" ] || printf '%s\n' "$NEW_IDS" | ssh "$@" "
umask 077 ;
mkdir -p .ssh && cat >> .ssh/authorized_keys || exit 1 ;
if type restorecon >/dev/null 2>&1 ; then restorecon -F .ssh .ssh/authorized_keys ; fi" \
|| exit 1
ADDED=$(printf '%s\n' "$NEW_IDS" | wc -l)
;;
esac
if [ "$DRY_RUN" ] ; then
cat <<-EOF
=-=-=-=-=-=-=-=
Would have added the following key(s):
$NEW_IDS
=-=-=-=-=-=-=-=
EOF
else else
if [ x$SSH_AUTH_SOCK != x ] && ssh-add -L >/dev/null 2>&1; then cat <<-EOF
GET_ID="$GET_ID ssh-add -L"
fi Number of key(s) added: $ADDED
Now try logging into the machine, with: "ssh $SSH_OPTS"
and check to make sure that only the key(s) you wanted were added.
EOF
fi fi
if [ -z "`eval $GET_ID`" ] && [ -r "${ID_FILE}" ] ; then # =-=-=-=
GET_ID="cat \"${ID_FILE}\""
fi
if [ -z "`eval $GET_ID`" ]; then
echo "$0: ERROR: No identities found" >&2
exit 1
fi
if [ "$#" -lt 1 ] || [ "$1" = "-h" ] || [ "$1" = "--help" ]; then
echo "Usage: $0 [-i [identity_file]] [user@]machine" >&2
exit 1
fi
# strip any trailing colon
host=`echo $1 | sed 's/:$//'`
{ eval "$GET_ID" ; } | ssh $host "umask 077; test -d ~/.ssh || mkdir ~/.ssh ; cat >> ~/.ssh/authorized_keys" || exit 1
cat <<EOF
Now try logging into the machine, with "ssh '$host'", and check in:
~/.ssh/authorized_keys
to make sure we haven't added extra keys that you weren't expecting.
EOF

View File

@ -1,75 +1,186 @@
.ig \" -*- nroff -*- .ig \" -*- nroff -*-
Copyright (c) 1999 Philip Hands Computing <http://www.hands.com/> Copyright (c) 1999-2013 hands.com Ltd. <http://hands.com/>
Permission is granted to make and distribute verbatim copies of Redistribution and use in source and binary forms, with or without
this manual provided the copyright notice and this permission notice modification, are permitted provided that the following conditions
are preserved on all copies. are met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
Permission is granted to copy and distribute modified versions of this THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
manual under the conditions for verbatim copying, provided that the IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
entire resulting derived work is distributed under the terms of a OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
permission notice identical to this one. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
Permission is granted to copy and distribute translations of this NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
manual into another language, under the above conditions for modified DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
versions, except that this permission notice may be included in THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
translations approved by the Free Software Foundation instead of in (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
the original English. THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
.. ..
.TH SSH-COPY-ID 1 "14 November 1999" "OpenSSH" .Dd $Mdocdate: June 17 2010 $
.SH NAME .Dt SSH-COPY-ID 1
ssh-copy-id \- install your public key in a remote machine's authorized_keys .Os
.SH SYNOPSIS .Sh NAME
.B ssh-copy-id [-i [identity_file]] .Nm ssh-copy-id
.I "[user@]machine" .Nd use locally available keys to authorise logins on a remote machine
.Sh SYNOPSIS
.Nm
.Op Fl n
.Op Fl i Op Ar identity_file
.Op Fl p Ar port
.Op Fl o Ar ssh_option
.Op Ar user Ns @ Ns
.Ar hostname
.Nm
.Fl h | Fl ?
.br .br
.SH DESCRIPTION .Sh DESCRIPTION
.BR ssh-copy-id .Nm
is a script that uses ssh to log into a remote machine and is a script that uses
append the indicated identity file to that machine's .Xr ssh 1
.B ~/.ssh/authorized_keys to log into a remote machine (presumably using a login password,
file. so password authentication should be enabled, unless you've done some
.PP clever use of multiple identities). It assembles a list of one or more
If the fingerprints (as described below) and tries to log in with each key, to
.B -i see if any of them are already installed (of course, if you are not using
option is given then the identity file (defaults to .Xr ssh-agent 1
.BR ~/.ssh/id_rsa.pub ) this may result in you being repeatedly prompted for pass-phrases).
is used, regardless of whether there are any keys in your It then assembles a list of those that failed to log in, and using ssh,
.BR ssh-agent . enables logins with those keys on the remote server. By default it adds
Otherwise, if this: the keys by appending them to the remote user's
.PP .Pa ~/.ssh/authorized_keys
.B " ssh-add -L" (creating the file, and directory, if necessary). It is also capable
.PP of detecting if the remote system is a NetScreen, and using its
provides any output, it uses that in preference to the identity file. .Ql set ssh pka-dsa key ...
.PP command instead.
If the .Pp
.B -i The options are as follows:
option is used, or the .Bl -tag -width Ds
.B ssh-add .It Fl i Ar identity_file
produced no output, then it uses the contents of the identity Use only the key(s) contained in
file. Once it has one or more fingerprints (by whatever means) it .Ar identity_file
uses ssh to append them to (rather than looking for identities via
.B ~/.ssh/authorized_keys .Xr ssh-add 1
on the remote machine (creating the file, and directory, if necessary.) or in the
.Ic default_ID_file ) .
.SH NOTES If the filename does not end in
This program does not modify the permissions of any .Pa .pub
pre-existing files or directories. Therefore, if the remote this is added. If the filename is omitted, the
.B sshd .Ic default_ID_file
has is used.
.B StrictModes .Pp
set in its Note that this can be used to ensure that the keys copied have the
configuration, then the user's home, comment one prefers and/or extra options applied, by ensuring that the
.B ~/.ssh key file has these set as preferred before the copy is attempted.
folder, and .It Fl n
.B ~/.ssh/authorized_keys do a dry-run. Instead of installing keys on the remote system simply
file may need to have group writability disabled manually, e.g. via prints the key(s) that would have been installed.
.It Fl h , Fl ?
.B " chmod go-w ~ ~/.ssh ~/.ssh/authorized_keys" Print Usage summary
.It Fl p Ar port , Fl o Ar ssh_option
on the remote machine. These two options are simply passed through untouched, along with their
argument, to allow one to set the port or other
.SH "SEE ALSO" .Xr ssh 1
.BR ssh (1), options, respectively.
.BR ssh-agent (1), .Pp
.BR sshd (8) Rather than specifying these as command line options, it is often better to use (per-host) settings in
.Xr ssh 1 Ns 's
configuration file:
.Xr ssh_config 5 .
.El
.Pp
Default behaviour without
.Fl i ,
is to check if
.Ql ssh-add -L
provides any output, and if so those keys are used. Note that this results in
the comment on the key being the filename that was given to
.Xr ssh-add 1
when the key was loaded into your
.Xr ssh-agent 1
rather than the comment contained in that file, which is a bit of a shame.
Otherwise, if
.Xr ssh-add 1
provides no keys contents of the
.Ic default_ID_file
will be used.
.Pp
The
.Ic default_ID_file
is the most recent file that matches:
.Pa ~/.ssh/id*.pub ,
(excluding those that match
.Pa ~/.ssh/*-cert.pub )
so if you create a key that is not the one you want
.Nm
to use, just use
.Xr touch 1
on your preferred key's
.Pa .pub
file to reinstate it as the most recent.
.Pp
.Sh EXAMPLES
If you have already installed keys from one system on a lot of remote
hosts, and you then create a new key, on a new client machine, say,
it can be difficult to keep track of which systems on which you've
installed the new key. One way of dealing with this is to load both
the new key and old key(s) into your
.Xr ssh-agent 1 .
Load the new key first, without the
.Fl c
option, then load one or more old keys into the agent, possibly by
ssh-ing to the client machine that has that old key, using the
.Fl A
option to allow agent forwarding:
.Pp
.D1 user@newclient$ ssh-add
.D1 user@newclient$ ssh -A old.client
.D1 user@oldl$ ssh-add -c
.D1 No ... prompt for pass-phrase ...
.D1 user@old$ logoff
.D1 user@newclient$ ssh someserver
.Pp
now, if the new key is installed on the server, you'll be allowed in
unprompted, whereas if you only have the old key(s) enabled, you'll be
asked for confirmation, which is your cue to log back out and run
.Pp
.D1 user@newclient$ ssh-copy-id -i someserver
.Pp
The reason you might want to specify the -i option in this case is to
ensure that the comment on the installed key is the one from the
.Pa .pub
file, rather than just the filename that was loaded into you agent.
It also ensures that only the id you intended is installed, rather than
all the keys that you have in your
.Xr ssh-agent 1 .
Of course, you can specify another id, or use the contents of the
.Xr ssh-agent 1
as you prefer.
.Pp
Having mentioned
.Xr ssh-add 1 Ns 's
.Fl c
option, you might consider using this whenever using agent forwarding
to avoid your key being hijacked, but it is much better to instead use
.Xr ssh 1 Ns 's
.Ar ProxyCommand
and
.Fl W
option,
to bounce through remote servers while always doing direct end-to-end
authentication. This way the middle hop(s) don't get access to your
.Xr ssh-agent 1 .
A web search for
.Ql ssh proxycommand nc
should prove enlightening (N.B. the modern approach is to use the
.Fl W
option, rather than
.Xr nc 1 ) .
.Sh "SEE ALSO"
.Xr ssh 1 ,
.Xr ssh-agent 1 ,
.Xr sshd 8