SSH with GDG new generation

pravle · New User Joined: 06 Jun 2006 Posts: 8

Hi,

I am trying to SSH to a GDG file from UNIX to Mainframe and it is not transferring it and giving error.
I gave the following command and as can be seen I am using the relative number i.e +1 here:
ssh -v username@remote.host "mv -P RECFM=FB,LRECL=150 /homedir/username/test.dat \"//'HLQ.XXXX.DATA(+1)'\" "

Error:
mv: FSUMF157 //'HLQ.XXXX.DATA' is an existing sequential data set, no member allowed

But when I try to use absolute number it is working i.e G0001V00 here:
ssh -v username@remote.host "mv -P RECFM=FB,LRECL=150 /homedir/username/test.dat \"//'HLQ.XXXX.DATA.G0001V00'\" "

The GDG index (base) is already built at the mainframe side.
Is there a way I can do SSH using relative number? The hard-coding will mean everytime I have to be aware of the last number and increment it manually.
The version of SSH is:
OpenSSH_4.3p2, OpenSSL 0.9.7d 17 Mar 2004

Thanks.

Robert Sample · Posted: Thu Mar 26, 2009 4:49 pm

I don't use OpenSSH here, but I suspect the error message is telling you that a GDG reference in relative form is not valid as used. Either plan on using absolute references, or change your processing so OpenSSH isn't attempting to send to a relative GDG reference.

dick scherrer · Posted: Thu Mar 26, 2009 9:04 pm

Hello,

You might consider sending the file as some constant non-gdg name and then copy it to the gdg+1 after the transmission. . .

pravle · New User Joined: 06 Jun 2006 Posts: 8

Thanks for the suggestions.

Yes, I am about to implement suggestion regarding sending ps file and then write a JCL to convert it into GDG.
But this requires creating extra JCL for every GDG that needs to be transferred. I was thinking that may be a solution will be there which would not require this.

dick scherrer · Posted: Fri Mar 27, 2009 10:30 am

Hello,

Depending on your standards, you could create a small proc to copy the file to a gdg that had symbolic parameters for the input and output dataset name (or some part of it).

Also, if you implement a separate copy step, i suggest your use the sort product on your system to do the copy. The sort products are exceptionally fast at creating a copy of a sequential file (much faster than iebgener or idcams).