psuedo file error during put with gdgnt option

Discussion of Co:Z sftp, a port of OpenSSH sftp for z/OS
Post Reply
dabills
Posts: 41
Joined: Thu May 19, 2011 9:56 am

psuedo file error during put with gdgnt option

Post by dabills »

We just upgraded to v2.4.1 from v2.3.2 and are getting the following error message when trying to PUT a GDG file:
"The put command does not support upload of a pseudo directory or non-leaf data set"

What was working in v2.3.2 isn't working in v2.4.1.

Here are the particulars that used to work in v2.3.2 but don't work anymore in v2.4.1:
lzopts mode=text,servercp=ISO8859-1,clientcp=IBM-1047,gdgnt
put //GDG.TEST.ED(+1) test.file
or
put //!GDG.TEST.ED(+1) test.file

will cause the failure message.

I was able to get this to work with v2.4.1 by doing one of the following:
1. remove gdgnt option
or
2. change to //GDG.TEST.ED(0)
or
3. change to //!GDG.TEST.ED(0)

Note: we do have a GDG file with the name GDG.TEST as well. It is unrelated to the GDG.TEST.ED file, but I suspect the GDG.TEST file may be causing part of the problem since it's named as part of the GDG.TEST.ED file name.

Did something change in v2.4.2 that is causing a change to the SFTP statements?
lisa
Posts: 5
Joined: Tue Jan 08, 2013 1:39 pm

Re: psuedo file error during put with gdgnt option

Post by lisa »

Yes, we did make significant changes in 2.4.1 in order to support more wild-card scenarios. These changes did affect GDG functions.
Please refer to http://dovetail.com/docs/cozinstall/changes.html for more information.

However, in the scenario that you have described, the put should have failed in 2.3.2 since it should be impossible
to read an existing generation with (+1) when the GDGNT option is enabled. This is a bug in Co:Z version 2.3.2.
The correct solution for your scenario is to remove the GDGNT option. With GDGNT set, the following error message should have been reported in 2.3.2:
IKJ56871I DATA SET dsnname NOT ALLOCATED, RELATIVE GENERATION NUMBER INCOMPATIBLE FOR SPECIFIED STATUS

In a future release we will consider improving the message reported by 2.4.1 for this scenario.
Post Reply