Re: FTP rename results in two files in IFS

The Created time stamps seem to confirm just what you suggested - two FTP
PUT commands

Thank you!



CRPence <CRPbottle@xxxxxxxxx>
Sent by: midrange-l-bounces@xxxxxxxxxxxx
01/26/2011 01:47 PM
Please respond to
Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>


To
midrange-l@xxxxxxxxxxxx
cc

Subject
Re: FTP rename results in two files in IFS






On 1/26/11 11:03 AM, Gary Thompson wrote:

<<SNIP>> the results are put in a .csv file and sent by FTP to a
directory under the IFS Root on our iSeries, resulting in hundreds of
small files a day.

We run a process every few minutes to read these .csv files and post
picking results files on our iSeries

To avoid processing a file before the FTP process completes, we use a
naming convention where the voice pick server does the FTP PUT using
a file named with a prefix of FTP_, and the iSeries post process
ignores any such files.

When the PUT command completes, the VP server re-names the file,
removing the prefix.

About 98% of the time (just a guess) the re-name works as expected.

For some files, however, the re-name results in two files, the
"first" file that has the FTP_ prefix and the "second" or re-named
file without the prefix.

Has anyone seen this?

Any suggestions as to a fix?

Might be worth investigating... Do both files have [the same or any]
data, or just one of the two? How different are the timestamps [create
and change] for the files? By suggesting the "VP server renames the
file removing the prefix", I infer that action is the effect of a line
in a prepared FTP script using the FTP subcommand RENAME such as the
unquoted "REN FTP_TEST TEST".? If a scripted request...

Probably then, two invocations [not necessarily concurrently] are in
conflict for having sent the same file name. On a second send of the
same file name, the rename request will fail with FTP550 and both file
names will persist. A REName failure is probably being ignored for
"file name already exists"; i.e. the script processing by FTP will log
the error, but there is no means within the scripting language for a
MONMSG or ON-ERROR. So unless a script which was generated to have the
FTP subcommand RENAME were then scrubbed for errors, that the rename
failed could easily be unknown.

Regards, Chuck