RE: Data Transfer

Don,

If you need help fixing your setup for SAVRSTLIB to resolve the Route to specified location not found, please contact me offline.
If you're only doing 1 library, and under 100 gb, SAVRSTLIB definitely way to go.
No need to worry about creating and clearing of the SAVF, both LPARS.

Paul

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of rob@xxxxxxxxx
Sent: Friday, September 06, 2013 2:18 PM
To: Midrange Systems Technical Discussion
Subject: Re: Data Transfer

I, too, have heard that transferring save files was significantly slower than stream file data.

http://pic.dhe.ibm.com/infocenter/iseries/v7r1m0/topic/rzam4/rzam4virtualtape.htm




Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1 Group Dekko Dept 1600 Mail to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com





From: CRPence <CRPbottle@xxxxxxxxx>
To: midrange-l@xxxxxxxxxxxx,
Date: 09/06/2013 01:29 PM
Subject: Re: Data Transfer
Sent by: midrange-l-bounces@xxxxxxxxxxxx



On 05 Sep 2013 07:59, Don Wereschuk wrote:

I'm having a problem transferring current data from a production box
to our development box. I've tried using FTP but I'm having issues
with this. First I did a SAVF on the library <<SNIP>>
100 GB of data to transfer. My problem is that in using FTP it is
taking about 3 hrs. to FTP 16 GB of data. This would result in a time
of almost 20 hrs. to transfer the 100 GB. Is there a faster way to do
this or is using a tape backup and physically delivering it to the
other machine the fastest way? I only have a window of approx. 4-5
hrs. Any and all suggestions are appreciated.

Writing to a Save File is considerably slower [at least according to
past discussions] than writing a binary stream or binary non-save-file
[i.e. database] fixed-length records. Thus switching from using SAVF to
using Virtual Tape files is probably a better\faster choice. Plus, the
choice of what to save in any one save request would no longer be
limited to one library per save as required when using save files.

I would suggest testing\timing a transfer of the 16GB save file data
again, but PUT the file data into a stream file. The timing of that
request would give a better idea of how much faster the transfer of the
same saved data could be effected using virtual tape images than when
using a save-file; i.e. the difference should imply the overhead of
writing save file records vs writing non-record\binary stream data.