Hi Partha Because of the problems nohup gives, I have stopped using it completely. Instead I use "screen". screen is much more reliable and the good thing about it is that you can monitor outputs remotely without them being written to a file. Plus you can even give inputs remotely or after logging out and logging in again.
screen gives you a shell, from that shell you can run your process (with an '&'). And then you can detach the process from the current login session by typing screen -d . Then if you logout the process continues to run. To get back to that shell, do screen -r . Look at man page of screen for more help. But I guess this may not completely solve your problem if the partial data you want to read does not fit a screen. A way to use partial outputs could be to write what you need to files from inside your code and close those files so that you can read them. In any case, screen provides more ease than nohup in finding the reasons of crashing of a process. I hope that helps. - Amit On Mon, 19 Jul 2004, Partha Konar wrote: > > Hi Amit, > I have also some problem using nohup at Solaris and I think > it is close to your one. > Previously I used DEC machine where nohup worked perfectly OK. > Now, in Solaris when I am running some process using > > > nohup a.out & > > It creates nohup.out at once, but does not write any thing before > gathering data of size 8192 (if the process is still running). > > If process is terminated before getting that amount of data, in both cases > whether you killed the process yourself or it is killed due to some bug > during run, nohup.out remains empty. > > Of course, if the process is killed when data size (s) is 8192 < s < 8192 * 2 > you will get only first bulk of data of size 8192. > > > ** Is there any way to get data inside nohup.out as soon as it create ? > This is the case in DEC machine. > > ** Then one can start using these data without waiting for the programme > to be finished ! > > ** Then the problems of loosing data if the process is killed are > automatically solved. > > Can you help me ? bye. > .. Partha Konar > > > --- The meme for blind faith secures its own perpetuation by the simple unconscious expedient of discouraging rational inquiry. - Richard Dawkins _______________________________________________ Solaris-Users mailing list [EMAIL PROTECTED] http://www.filibeto.org/mailman/listinfo/solaris-users
