On Mon, Oct 15, 2012 at 10:04:39AM +0800, Tinni wrote: > Thank you so much for your reply and explanation. > > >>So, you're trying to retrieve 3 separate (string?) values from a remote > system. > > Yes. > > >> Why not simply prompt the user for the > information on the local system and skip the ssh call altogether? > > Because the script has to run in database servers.. > > > In your explanation , if I understand correctly, it will take the parameter > value from test1.sh in the remote server and returned to the local server > right? > > > >>source ~oracle/.whatever || exit > > Can I use source ~test1/test1.sh to call into remote server?
If test1.sh is the one that has the embedded terminal escape sequences and the prompts and the read commands, then you're not going to be able to use it in this way. All of the output it produces (the prompts and the terminal escape sequences and the labels and status indicators and so on) will contaminate the data stream that you're sending back to the local system. The interactive "read" commands will attempt to pull data from stdin, but in my example I was using stdin to send the commands to the remote bash shell, so you can't use it to read from the user's keyboard. If you truly need to prompt the user for information and then use that information to communicate with a database to get DIFFERENT information, all of it across an ssh session, then you have a monstrous set of problems to solve. > - To execute the script , first I have created /lhome/test1/.test in the > remote server assigning > the value in each parameter. But no values returned into the local > server. > > - Then I used /lhome/test1/test1.sh [ which was my original script in > remote ] ,, but it > did not wait for value to type , and as a result no value returned into > the local server.. > > Am I doing anything wrong? If it's not working then obviously you're doing something wrong. This problem is so difficult to solve that it's going to take you a much greater effort than I think you were expecting. > As I mentioned in my original email, that i need to execute the test1.sh > script in the remote server > and process, then the value i need in the local server for further process. But the test1.sh script is *not suitable* for this problem. You need to find another way to get the information, without using test1.sh. Take it apart, see what it does, and then see how you can perform those same steps without all the noise and interaction. > If my approach is not right, then would you please suggest , in this > situation, what kind of > approach I need to take to write the script!! Start by specifying exactly what results you need. Always keep this end goal in mind. Next, understand the working environment. You're trying to retrieve data from a remote system accessible with ssh, which is very different from a system where you have an interactive terminal with a shell in it, or a graphical interface. There are limitations in the ssh session that I don't think you've fully come to terms with yet. You may not be able to use the tools you're accustomed to using (such as terminal-based interactive shell scripts like "test1.sh") because they may not perform correctly in the environment you're working under. Then the final step is to figure out how to get the results you need in your working environment. This will require the proper selection and use of many different tools. You'll have to bring several techniques together in precisely the right way. I suggest that the proper venue for continuing this discussion is the help-b...@gnu.org mailing list, rather than bug-bash or private email.
