UNIX for Dummies Questions & Answers

The operating system is Solaris. There is a perl interpreter that is located at /opt/perl5.10.0/perl. when i give
ls -ltr /opt/perl5.10.0/perl
I get an error message saying that "/opt/perl5.10.0/perl Not a directory"
At the next instant when i give the same command it lists the properties of the file. For the next 12 minutes or so ls -ltr /opt/perl5.10.0/perl produces no error message. After 12 minutes, the server seems to forget the path /opt/perl5.10.0/perl [i.e when i give ls -ltr /opt/perl5.10.0/perl after 12 minutes it gives an error message Not a directory.]

I have a perl script that is run from autosys. This script makes use of the above perl interpreter. The job fails for the first time and then runs for second time and goes to success for the above mentioned reason. Can someone please help me find out the cause of the problem and shed some light on the solution.

This problem is quite funny and I have not come across this error message for ls command when i googled.

I don't know how autosys is interfaced with the system nor whether or not it calls routine systems about schedule related things.
But i know that i've already met a bug that was making the cron not 100% reliable.

One of you monitoring scheduled script that sould be run every n minutes was running every n minutes, but some randomized times, it was just missing a run (this by forcing the script to generate a log file every run that i located this issue.

That was a solaris 10 bug that got fixed patching with last patch update.

Don't know whether it has something to with you prob but it might be something to check... if you run this OS.

otherwise check environment settings at time you are running your script as well as the availibility of the directory.
Is this dir synchronized with another one ? any NFS or automount stuff in this tree ? no inode table issue ?
truss your failling command and see if you find something in the log

I suspect this error might be caused by /opt/perl5.10.0 being a link or a mountpoint to a remote file server and the latter service being faulty.

The error message suggests that /opt/perl5.10.0/perl is the current Shell.
If your script starts with a shebang line for perl, any error messages will be attributes to that instance.

We assume that /opt/perl5.10.0/perl is a file not a directory because you describe it as a file.

Assuming that it is a file, you can generate the error message with commands such as "cd /opt/perl5.10.0/perl" .

What is in your script?

hi sreeharshasn

looks like /opt/perl5.10.0/perl is on mount point which is mounted using autofs
which drops the mount point after certain period of idle time.

when u do ls -lrt /opt/perl5.10.0/perl at that time it does not exist it needs to be mounted first .. so for the first time ls -lrt returns error but that mounts that directory .. so that mount will work for some time ( that time is decided in auto mount )

if you want to make sure ( below commands runs on Linux )

Code:

$ cd /opt ; df . 
check the output of above command .. if auto mounted you will see autofs and PID in bracket  

$ service autofs status 
check if /opt or /opt/perl5.10.0 is listed there ..it will also list timeout count in seconds 

if not on linux, run below command 

$ /etc/init.d/autofs status

i have faced same issue in which particular directory was not accessible first time you use it after reboot ..

hope this solves your problem

Thanks everyone for helping me understand the situation. The perl interpreter was actually mounted when the first time an ls command was given.