Text 1268, 97 rader
Skriven 2009-05-06 08:00:06 av Mike Luther (1:117/3001.0)
   Kommentar till text 1180 av Sean Dennis (1:18/200)
Ärende: Synchronizing files
===========================
Hi Sean .. and (All here)

 SD> Hello, All.

 SD> Is there a good simple (and free!) program to allow me 
 SD> to synchronize two directories on two different 
 SD> drives?  I'm wanting to do that on the fly with my 
 SD> Cheepware development directory to back it up to a ZIP 
 SD> 100 drive.

Peter Knapper popped a reply to me in the OS/2 Storage NewsGroup that had the
magic word in it.  We are talking about what is documented in the OS2
Assistance packaging about File Replication.

     1.) Open your Assistamce folder on your OS/2 MCP2 Desktop
     2.) Open the Tasks subfolder object.
     3.) Open the File and Print Client GUide subfolder object.
     4.) Look for the File Replication header - expand it.

     Be prepared for a VERY intensive study session on how to do all
     this. It even reqires the use of a text editor to prepare the
     control files for getting all this done.

I've never used this at all.  From reading the documentation, I think this is
what is being sought here but the control process requires a real understanding
of how the whole OS/2 toolset for it works and the use of a complete manual
setup proceedure with a text editor to enable.

From what I recall Peter's coaching in the old pointers he posted at this a
long time ago, the replication is actually a chunk by chunk file duplication
process in the OS/2 LAN operations, essentially a pre-cursor sort of process to
what is like the RAID operations of today.  If what I recall of this is
correct, if you enable this, you will get a step-by-step mirroring of the
workstation file(s)/directory(ies) in real-time as the master file is modified
by whatever is accessing it for 'normal' use.  That from the LAN/PEERLAN server
where the master file is being stored and accessed.  Accessed ... hmmmmmmmmmm. 
;)

However .. never having used this toolset, I have no idea what issues one might
face with using it as a mission critical issue while the server is also being
used for, say, multiple user access on LAN/PEERLAN for shared files by many
users with even carefully orchestrated and designed locked access sequential
write modification for given shared record chunk access for any user during the
whole multi-user open for read master access for this work.

That, for instance, with what might happen if the server were simultaneously
being used for heavy multi-user COMM port and/or TCP/IP service as well,
especially if the protocol NETBIOS over TCP/IP is installed on the OS/2
LAN/PEERLAN units in use.

For example, what if you are using the Btrieve mission critical file and
indexing toolset for either DOS or OS/2 network multiple user or multiple
workstation/user operations, together with COMM port and other TCP/IP
operations in different applications at the same time?  Such as running your
BBS in one operation for POTS, another one for TelNet,another one for another
node, an FTP server, HyperAccess HHOST, and total phone line monitoring of,
say, six POTS Lines with COMM port constant I/O all at the same time all your
multiple Btrieve files on the same server were bing networked to a host of
other workstations in real-time?

Which is exactly what I do with all my stuff here with all of my own personal
created operations and code.

See why I was sometimes such a pest on IBM's TestCase, chuckle?  And how it
would be a horrible error tracing issue to fix what could go wrong in all the
ring level code that is so brilliantly written into OS/2 at times?  That gets
broken by OS/2 as a result of COMM port or TCP/IP operations when somehow the
cows are running down the chute in real-time and something else yanks the chute
out from the trailer and the holding pen below?   Moooooooooooo .... moan.

Looking back at this, I now realize how badly I should have tried this on the
work for all those years on TestCase with File Replication which is part of
OS/2 and how I failed us all by not doing that in my work.

As well as missed a huge opportunity to perhaps open up the process of mission
critical file chunk by chunk replication on a workstation that could have
seriously simplified the whole backup operation of absolutely mission critical
files and directories without RAID.  I'm sorry I didn't understand enough about
the OS/2 opportunities to do this folks.

But at least the tool is here if someone wants to try it.  And hopefully our
wedding ring we all get for help with that won't break ring zero or another
more important but lower level relationship...

     Chortle.


--> Sleep well; OS/2's still awake! ;)

Mike @ 1:117/3001





--- Maximus/2 3.01
 * Origin: Ziplog Public Port (1:117/3001)