View previous topic :: View next topic |
Author |
Message |
CZerfas Intermediate
Joined: 31 Jan 2003 Posts: 211 Topics: 8
|
Posted: Wed Apr 15, 2009 3:11 am Post subject: Counting Bytes of a dataset |
|
|
To ensure that a transmission of datasets from one data center to another via FTP has completed correctly, I want to count the numbers of transfered members (if it is a PO dataset) and the number of used bytes.
This last bit of information seems not easy to find. The smallest intervall to be countable seems to be the allocation unit, but the requirement for the content of my report is the number of bytes.
What is your oppinion: Is the number of allocation units used sufficient according to my requirements? Is there a way to actually count bytes?
with many thanks
Christian |
|
Back to top |
|
|
CZerfas Intermediate
Joined: 31 Jan 2003 Posts: 211 Topics: 8
|
Posted: Wed Apr 15, 2009 6:15 am Post subject: |
|
|
To make things clearer: I need a solution for sequential datasets, PO-datasets and load libraries.
Thanks
Christian |
|
Back to top |
|
|
dbzTHEdinosauer Supermod
Joined: 20 Oct 2006 Posts: 1411 Topics: 26 Location: germany
|
Posted: Wed Apr 15, 2009 7:08 am Post subject: |
|
|
instead of writing my own rules about what constitutes successful completion,
I would do a little research to determine what rules 'convening authorities'
(I don't mean the bean counters in your company)
have established to 'prove' a valid transfer of data has been accomplished.
this is not a new-technology question.
data transfer has been going on for a couple of months now.
this reminds me of companies establishing their own edi formats:
they had invested time and money into creating their own
and had no desire to spend more money to conform with the rest of the world.
apologize for the sermon. _________________ Dick Brenholtz
American living in Varel, Germany |
|
Back to top |
|
|
DaveyC Moderator
Joined: 02 Dec 2002 Posts: 151 Topics: 3 Location: Perth, Western Australia
|
Posted: Mon May 18, 2009 3:01 am Post subject: |
|
|
And how do you transfer PO datasets via FTP? I would just use AMATERSE which unloads the dataset to a sequential file and transfer that. It's got CRC checking - no worries.
Counting bytes is a useless scheme. _________________ Dave Crayford |
|
Back to top |
|
|
taltyman JCL Forum Moderator
Joined: 02 Dec 2002 Posts: 310 Topics: 8 Location: Texas
|
Posted: Mon May 18, 2009 8:08 am Post subject: |
|
|
Doesn't ftp tell you?
Code: | 250 Transfer completed successfully.
729 bytes transferred in 0.050 seconds. Transfer rate 14.58 Kbytes/sec. |
Send your output to the next step and parse it. |
|
Back to top |
|
|
DaveyC Moderator
Joined: 02 Dec 2002 Posts: 151 Topics: 3 Location: Perth, Western Australia
|
Posted: Tue May 19, 2009 8:46 am Post subject: |
|
|
Is this thread a bit silly. If you transfer a data set via FTP and there was a failure FTP will return with a > 0 return code. Shouldn't you be checking the return code from FTP? _________________ Dave Crayford |
|
Back to top |
|
|
CZerfas Intermediate
Joined: 31 Jan 2003 Posts: 211 Topics: 8
|
Posted: Wed May 20, 2009 9:05 am Post subject: |
|
|
Hello all,
after this thread has been waiting a while for answers, I implemented the following solution:
- with a TRANSMIT command I converted every dataset into a sequential FB80 file
- with TRANSMIT and RECEIVE these datasets are copied to the receiving LPAR
- with (local) RECEIVEs into other DS-NAMEs the original structures of these files ares restored on the receiving LPAR
With this strategy all file attributes are restored and the (local) RECEIVE does CRC checking, therefore the resulting dataset can be considered as "correct".
I simply added a counting of the members per dataset on the sending and the receiving LPAR, and than even the managers have been satisfied.
Just out of curiosity I tried to send regular FB80 libraries via FTP to the receiving LPAR, and we often had corrupted members and even corrupted files. Therefore, a FTP transmission ending with a RC=0 can't be considered to be "correct" always.
regards
Christian |
|
Back to top |
|
|
DaveyC Moderator
Joined: 02 Dec 2002 Posts: 151 Topics: 3 Location: Perth, Western Australia
|
Posted: Thu May 21, 2009 10:07 am Post subject: |
|
|
CZerfas wrote: |
Just out of curiosity I tried to send regular FB80 libraries via FTP to the receiving LPAR, and we often had corrupted members and even corrupted files. Therefore, a FTP transmission ending with a RC=0 can't be considered to be "correct" always.
|
Then you're usage is incorrect or IBM have a very serious bug! Did you use the MODE command? The normal method is a AMATERSE step and then FTP. Much better than XMIT and now a supported utility. _________________ Dave Crayford |
|
Back to top |
|
|
|
|