FTP session from V7R1 -> V5R4

[Mr. K.V.B.L.]

Sent a *SAVF from V7R1 to a V5R4 machine. The *SAVF was sent
successfully as I was able to display on the target machine but the
source was sitting there, like it wasn't quite done. After a while I
guess it timed out. We're working out some networking trouble on the
source machine and I was wondering if this is a problem we still need
to look at or is this a problem from going to one machine to another
that is on a different version of the system. Notice that the source
machine never said that the file was successfully sent, only sending
then the timeout message, but it did successfully get there.

Connecting to remote host XXX.XXX.XXX.XXX using port
21.
220-QTCP at
LEGACY.
220 Connection will close if idle more than 60
minutes.

>
331 Enter
password.
230 BEAK logged
on.
OS/400 is the remote operating system. The TCP/IP version is
"V5R4M0".
250 Now using naming format
"0".
257 "BEAK" is current
library.
>
bin
200 Representation type is binary
IMAGE.
> lcd
beakstuff
Local working directory is
BEAKSTUFF
> cd
beakstuff
250 "BEAKSTUFF" is current
library.
> put df990004.savf
250 Now using naming format
"0".
257 "BEAKSTUFF" is current
library.
500 Subcommand EPSV not
valid.
227 Entering Passive Mode
(50,73,82,40,30,250).
150 Sending file to member DF990004 in file DF990004 in library
BEAKSTUFF.
No response from remote host; all connections
closed.
Enter an FTP
subcommand.

[JHL]

PTFs up to date on both boxes?
remember you must save to the lower target release if you may want to
restore the savf at 5.4

joe www.michsoft.com

[CRPence]

On 20-Feb-2012 09:55 , Mr. K.V.B.L. wrote:
> Sent a *SAVF from V7R1 to a V5R4 machine. The *SAVF was sent
> successfully as I was able to display on the target machine

What does "was able to display" mean? That DSPSAVF was successful?
If so, that only indicates that the server dropped the lock on the save
file and that at least the records which record what the later data
should be. That does not indicate all of the data is there; i.e. like
tape, a save file is sequential media whereby earlier data records what
should follow... not what actually follows, which is only known after
the following data is read.

> but the source was sitting there, like it wasn't quite done. After a
> while I guess it timed out.

So although the server dropped the lock, possibly after receiving an
indication all available data was sent and received, the client
continued awaiting some handshaking to confirm the completion.

> We're working out some networking trouble on the source machine and I
> was wondering if this is a problem we still need to look at or is
> this a problem from going to one machine to another that is on a
> different version of the system.

Binary data is irrespective of release. The FTP is merely
transporting data, regardless any implication by the acronym.

Look for discussions about the FTP500 "Subcommand EPSV not
valid." message, IIRC, regarding some changes that may need to be
accommodated due to changes with the v7r1 [IBM i 7.1] FTP. IIRC some
past discussion made reference to the MTU for 7.1, implying the EPSV was
discussed there.

> Notice that the source machine never said that the file was
> successfully sent, only sending then the timeout message, but it
> did successfully get there.

Without a successful completion message, I would not consider the
data to have been "sent successfully"; except possibly by inference from
the same number of records in the file on the source and target, and the
ability both to restore from the data in the save file without error and
the ability to access any restored objects and the data within. Of
course a binary compare of the stream data or record data of the save
files would be even more thorough, but that may in effect, require a
successful data transport via FTP or another mechanism.

Regards, Chuck

[Mr. K.V.B.L.]

The data was restored to another library successfully on the V5R4,
FWIW. The save file was originally created on V5R4, sent to V7R1,
then sent back. The V7R1 file says it's release level is V5R4. I see
this as a protocol problem (not likely) or a networking problem. I'm
going to do the test again to see if it's predictably unreliable.

[Jonathan Bailey]

There are 2 timeout settings on the as400. If you enter
quote stat
At the end (at 5.4) I get
211-File transfer time-out value set to 420 seconds.
211 Current inactivity time-out value set to 300 seconds.

So if I started a file transfer which took more than 420 seconds I
think the system would end the send session abnormally, If the data
was sent in less than 300 seconds I could then enter another command.
If it was sent between these values the data send should complete but
the system would log me off due to inactivity.

try quote help time followed by quote time 10 5000
then send a 'massive' file before the 10 seconds are up then get back
to us with your findings.

HTH

Jonathan

> going to do the test again to see if it's predictably unreliable.- Hide quoted text -
>
> - Show quoted text -

[Franco Lombardo]

Maybe this IBM's doc could help you:

http://bit.ly/wZCOzc

Bye

Franco


--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
http://www.francolombardo.net
Scala, Java, As400.....
http://twitter.com/f_lombardo
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~