Help with net-ssh shell-script timing

ruby-talk
1

I'm working on a script to connect to a remote server, run a custom su
script, and then execute commands after it finishes and hands the root
prompt back.
my main issue seems to be that net-ssh likes to fire-and-forget commands
sent over the channel. what's killing me is this sudo script, i cannot
alter it, and i cannot get to the files i need to output without going
through it first.

···

___
the script behaves like this
user@blah

ss

Password:

executing script blablabla...

root@blah

___

what i have so far:

SSH.start(host, user){|ssh|
    ssh.open_channel{|channel| #get root privelages
        #configure behavior of channel
        channel.on_data{|channel, data| #when output is recieved from
remote stdout do...
                puts "#{data}"
                if data =~ /^Password:/
                   channel.send_data("#{PASSWORD}\n")
                else end
        channel.on_close... (etc.) channel.request_pty do |ch,
success>
                if success
                puts "pty successfully obtained"
                else
                puts "could not obtain pty"
                end end

        channel.exec("ss"){|channel, win| #custom sudo script.
                if win
                   puts "ss command sent"
                else puts "ss command FAIL"
                end
        }
        channel.exec("tail /some/log/file.txt")
        #or channel.send_data
        # ssh.exec, and exec!
...end

what i've tried so far (that i can remember)
i've tried to block the exec("ss") command,
syncronizing exec"ss" the .on_data password callback, and the tail
command.
tried having the channel wait() after calling SS,
nesting the tail command in the on_data callback right after sending the
password.
tried having the on_data callback update its self something like this...
   channel.on_data{|channel, data|
           if data =~ /^Pssword:/
              channel.send_data("#{PASSWORD}\n"){
                channel.on_data{|channel, data| #executing script
blablabla...
                  channel.on_data{|channel, data|
                    channel.exec("tail /some/log/file.txt")
                  }
                 }
                }

i've come at the problem from every angle i can come up with, and still
need to figure out how ensure that if SS returns a root prompt without
requesting the password (if i've logged in recently) that the script
will proceed.

you all have been phenomenally helpful in the past, thanks in advance
for any assistance
~Locky

--
Posted via http://www.ruby-forum.com/\.

2

update: i've gotten a bit further, i can now at least get the program
to spit out the data i need... but it does it in an infinite loop >_>

SSH.start(host, user){|ssh|
    ssh.open_channel{|channel| #get root privelages
        #configure behavior of channel
        channel.on_data{|channel, data|
                puts "#{data}"
                if data =~ /^Password:/
                   channel.send_data("#{PASSWORD}\n")
                elsif data =~ /root@/
                    channel.exec("tail /some/log/file.txt")
                    channel.on_data{"STOP LOOPING, DAMN YOU!"}
                end
        channel.on_close... (etc.)

        channel.request_pty do |ch,success|
                if success
                puts "pty successfully obtained"
                else
                puts "could not obtain pty"
                end end

        channel.exec("sudoshell"){|channel, win| #custom sudo script.
                if win
                   puts "ss command sent"
                else puts "ss command FAIL"
                end
        }
...end

if i put the 2nd command anywhere but within an on_data if clause, it
refuses to wait for the sudoshell to finish loading, all attempts to use
a sleep, or wait command just result in it hanging forever.

could really use any advice on this one. thanks
~Locky

···

--
Posted via http://www.ruby-forum.com/.