Weird rcp behavior

[Christopher T. Pestano]

Hi,

I'm copying 4GB files between 2 servers using rcp. When I manually copy, the 4GB file is copied correctly but when use it inside a script and schedule it to run using at, it only copies 2GB of data. Any ideas why? Any help will be greatly appreciated.

Largefiles is enabled on both filesystems.

Regards,
Tope

[Steven E. Protter]

I have seen similar, strange behavior from rcp in the past. This is one of the reasons I dropped the r-protocols in favor of secure shell.

To help further, I'd need to know your OS version and patch level.

For test purposes, you might want to try the transfer with NFS.

I would check patches for these products.

SEP

[Christopher T. Pestano]

Thanks Steven,

You gave me an idea, maybe some missing patches. I'm running 11.11 on Superdome, with GoldBASE Dec2001 on the problematic servers. Havent' had the chance to update the server for a long time already. I have another machine which is not experiencing that problem which has QPK June2003. Hopefully a patch will resolve this, will browse through the patches now.

Cheers,
Tope

[T G Manikandan]

NFS is not the right choice as nfsv2 does not support largefiles.
Rcp is also the way to copy the files.

doing a ftp for the file transfer would be a better option.
Also splitting the files and then transfering would be other option.


I am not sure about the weird problem above but what user you are trying to transfer the files in the script.
check the limit for the user using ulimit.


Revert

[Patrick Wallek]

Try this -- type 'ulimit -a', assuming /usr/bin/sh or /sbin/sh (POSIX shell) and see what your "file (blocks)" limit is. Now put that line into your 'at' job script and see what it returns. It could be that from within 'at' you have a 2GB limit, but from the interactive shell it is unlimited.

If that is your problem, you can try putting the statement 'ulimit unlimited' into your script and see if that helps.

Good luck.