On Tue, 4 Apr 2017 00:03:43 +0000, Rob Schramm wrote:
>crlf lf nl comes to mind when going between platforms. But I like John's
>suggestion about checksum.
>
o ASCII transfer of a CSV file between any two of Windows, Linux, and z/OS
is almost certain to invalidate a checksum (cr/lf/nl).
o Checksums are uncommon among Windows platforms, but: Cygwin or
https://support.microsoft.com/en-us/help/889768/how-to-compute-the-md5-or-sha-1-cryptographic-hash-values-for-a-file
o And the customer user may be averse to the need
But if the CSV arrives intact at the "portal" (does it?), checksums need
only be calculated from that point on.
What OS and software is the "portal" using?
(Cue Timothy on the advantages of running a server ("portal") on the mainframe.
The fewer steps, the better.)
Is there any distinguishing characteristic of the lost records (length, content,
position in file, ...)?
Can you reproduce the problem by emulating the customer's behavior?
>On Mon, Apr 3, 2017, 7:34 PM John McKown wrote:
>
>> On Mon, Apr 3, 2017 at 5:15 PM, Ron Thomas wrote:
>>
>> > Hi . We have customer users uploading sales data in excel format in a
>> > portal, once this file is uploaded a batch process will SFTP the file from
>> > here to a virtual machine and then to the mainframe . We are seeing that
>> > there are many cases where all the rows the user entered in excel file is
>> > not getting transmitted to virtual machine .
>> >
>> > How we can make sure that all the rows gets transmitted to virtual machine,
> > > if all the rows gets transmitted then only transmit to mainframe else
>> > is there any specific design considerations we need to follow ?
>> > Please suggest valuable thoughts on this ?
>>
>> So, what is the customer using to send the file? It appears that this is
>> what is failing. So that is what needs to be detected. And, at this point,
>> I can't come up with anything better since I don't know what is being used.
>> It might be upload via a web page, or ftp, or sftp, or ??? . In a general
>> sense, I would say that the originating system needs to create a "checksum"
>> of the data and upload both that and the data. The receiving system, linux
>> on the VM, needs to regenerate the checksum and compare it to the
>> transmitted checksum. If they are different, then take corrective action. I
>> don't know what that would be.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
