product for mapping UNIX box

[larryt]

Task: map the UNIX box , due to allow Acu client products to "see-and-use" the COBOL sources, copy libs, etc. ... located on a UNIX server. What product should be taken in consideration? Samba ? Some other ones ?

I would appreciate your recommendation based on your experience with it.

[vinsnash]

If you're using SCO it comes packaged with 'Vision FS'. I've worked on systems using 'Vision FS' and 'Sama' and both have been more or less transparent from an Acucobol user's perspective.

I've used both in development environments and some limited applications that were not I/O intensive avoiding the necessity of Acuserver or Thin Client.

[dlitwin]

Here at Acucorp, we use Samba 'to "see-and-use" the COBOL sources, copy libs, etc. ... located on a UNIX server'.

For instance, you can have your sources & copylibs stored on a UNIX box with Samba, and still have AcuBench find them, and work with them.

However, the use of Samba mapped drives is not supported with regard to data files and a running COBOL program.

[golfs4us]

For years we have used a product called FACETWIN from facetcorp. http://www.facetcorp.com

We have been extremely pleased with the product and the support.

[MerlinM]

I assume you mean you are using win32 based clients with a linux/unix based server. Since you mentioned source code, I strongly advise (even urge!) you to become familiar with cvs to hold your source code repository. cvs works very well with AcuCobol and will increase your productivity (even if your build environment is on the unix server). As for other files (like data files, etc.) as the others said there are a lot of helpful utilities...with win32 clients samba is probably the best choice.

Merlin

[larryt]

I've just got from summer vacation and did a quick look at the sugested solutions.
Thanks all for sharing your experience.
We (developers) are using NCR UNIX with WIN clients. Unfortunately it seams that FACETWIN is not supporting NCR version of the UNIX.
On the other hand if I understand you correctly, Samba does not allow a developer using AcuBench to RUN programs located at the UNIX server? It is OK only for accessing sources, copys, ... but not for actual run & debug on the server side?
I am not sure what "cvs" stands for?
Thanks again for your advises.

[MerlinM]

Samba is a program which (among other things) allows unix mounts to be set up as a windows file share. You click on network neighboorhood, and browse your unix directories as if it was a windows machine. Samba is an SMB server. It is very fast.

'cvs' stands for concurrent versioning system. It is a source control repository (think: database with version control) that allows multiple developers to work on the same source files.
In cvs, each developer has a full copy of the entire source of the project that is kept synchronized with the server's copy. Compilation is done on the developer's machine, so each developer has to have a (licnesed) copy of the compiler that runs on his or her machine. (thus, each developer needs a licensed copy of ccbl32.exe for windows compilation).

Both cvs and samba come with most versions of linux...not sure about ncr unix.

When you say 'server side' run and debug, what is your client server architecture? Is the application process running on the client, or the server?

Merlin

[dlitwin]

To clarify my earlier warning about using a Samba-enabled mapped drive in the production environment:

You MAY locate your ACUCOBOL-GT runtime, or any compiled program objects on a Samba-enabled mapped drive. The Windows OS will load and run the wrun32 executable if stored on a Samba type drive. And the wrun32 runtime will load and run compiled objects stored on Samba type drives.

What is NOT supported:

Samba is not supported by Acucorp for DATA files, because locking operations are not always performed as expected. This can cause file corruption problems. You may find that Samba works for a while. But, then you may have problems when user activity reaches a certain level. It's risky and completely unpredictable. If you want to locate your Vision data files on the Unix server and access them from the Windows platform, we suggest our AcuServer product.


To summarize: Just don't locate your data files on a Samba drive.

[MerlinM]

Originally posted by dlitwin

To summarize: Just don't locate your data files on a Samba drive.

That is good to know...We use AcuServer here and are very happy with it. I also suspect that the AcuCorp programmers use cvs becuase of the RCS tags in some of the sample files :)

Acuserver is an order of magnitude faster than network shares, plus you get the administrative benefits...(boot user, etc.) it's really worth the cash.

Merlin