Backup software

[lionel0201]

We need a good backup/restore utility for eCS. It needs to allow backup & restore to internal/external hard drive, and must support JFS and files greater than 2 GB.

It does not need to be free! I would be willing to pay upward to US $100 for such supported application.

Thanks for listening

Lionel Abrahams

[troll]

I can't use eCs today because apps are missing.
When I used OS2/eCs actively, I used DualStor by IBM as backup, this is the only backup routine I have seen that really works and has an simple gui. DualStor is working also today in eCs, problem is emergency restore does not work due to some dlls and many new backupmedia is not supported. I'm quite focused on backup, yet today I have not seen a backup program/rutine as good as DualStor. I think it's awayable at hoobs and might be it's possible to get a deal with IBM so someone can modernize it for eCs? It's small, just an floppy, it works, rexx scripting, easy to use, few bugs. Compared to backup programs for windos to day, DualStor is superior, as OS2/eCs still is.

[warpcafe]

Hi all,

very intersting topic. By the way, IBM's DualStor also was my preferred backup solution (some years ago) because it was very easy to use and "nice". The bad thing was that (back in that time) it was already hard to find suitable tape drives (or suited drivers)...

Now time has changed. The standard user doesn't seem to have SCSI tape drives in most cases and people are shifting to use CD/DVD or just USB/MSD or simple "hard disk" drives for backup. This makes it easy to come up with even a rexx-based solution. Indeed, it is something I am intersted in a) using and b) perhaps developing.
Before you get too excited, here is some thoughts around that topic:

While most of the source/target and file/directory selection and processing is pretty "easy", the hard part is in details with EAs and the local "catalog" management. Configuration and scheduling with backup methodologies (incremnetal, full,... etc.) can be solved too. The main "hard" things in my mind is: Restore (not even to speak of disaster recovery).
All this can of course be determined and solved. We just need to sit together and think about requirements.

What's your wishlist? Ideas? Suggestions?
Cheers,
Thomas

[troll]

I'm glad someone has the same experience as I.
As said before, DualStor works also today with old media. Much old apps for OS2 works as DualStor but not fully because of old DLLs and drivers. it should be easy to modernize them to functional software. I'm not for doing things from scratch if not necessary. I'm not a programmer, but I can understand that they will have more fun doing it from scratch. But I don't thing this is the time. So they who have the knowledge and expertise has to listen to us who are users - What do we need and what is easy and fast to get up running up to date.

Functional Backup is one of the most important apps to get working first.

[warpcafe]

Hi,

no, I'm not talking about doing it from scratch just because it is more fun:
The problem simply is that you need access to the source code for maintaining a program/driver/dll. As far as I know (and by judging from the failure of various petitions to IBM) we do not have the source codes. So there is NOTHING you can do except doing it all over again.
It could be possible to only create a DLL from scratch, yes, but then you would need to know the parts that are in and are used by the program accessing it. And you would need to make it (the new DLL) 100% "interface compatible", else the application will not work in the intended way... if at all.

So it's not just to please me. I am afraid we don't have any other chance than doing something new.
On the other hand: Even if possible (which isn't), an "updated DLL" will give you the same DualStor. No new features (does it support files > 2GB? JFS?), no fixed bugs (if any). Doing it from scratch could be a chance for improving - making it faster, better, more pretty... whatever. It could be added more languages and we would have complete control over what (and how) we do as disaster recovery...

Cheers,
Thomas

[troll]

Don't you think IBM will release necessary information or do you know?

[warpcafe]

Hi,

ehh.. well, if I would ask IBM about releasing DualStor sources and DLL sources to the public, this is what would happen:
First, it would take them around 2 weeks to find someone who knows what the question is about.
That person would try to see if they still have the source code somewhere.
Then it's lawyers time to check if
- DualStor was sold to someone else
- DualStor and the DLLs use 3rd party components or have been made by a 3rd party
- there is any kind of copyright infringements or legal risks

Then they will calculate what it takes (time,money,skills) to change the IP agreements for DualStor to -let's say- GPL.
Then they will check if any other software using these DLLs can be impacted to their disadvantage.

If all that goes well, the might release it. But:
Look at all the steps mentioned and imagine the effort that it takes (time, money, people, skills,...) Do you think IBM is willing to take all these efforst in order to make something available "for free" (meaning: IBM gets no money for this).

And this is only under the preconditions that they know what DualStor is and have people with the skills available. If there is a chance of IBM having to hire contractors for investigating this - forget it. Too expensive.

Cheers
Thomas

[Alan Beagley]

very intersting topic. By the way, IBM's DualStor also was my preferred backup solution (some years ago) because it was very easy to use and "nice". The bad thing was that (back in that time) it was already hard to find suitable tape drives (or suited drivers)...

I used to use DualStor with an old Conner floppy-controller-connected tape drive. When I decided I had too much data to backup for their 120MB capacity and looked into the media cost of Travan, it was plain that it was too high and that the number of cartridge changes would still be inconvenient.

That's when I bought a refurbished SCSI DDS-2 autoloader (I was already using a SCSI hard disk) and switched to BackAgain/2. I am now using a DDS-4 autoloader (up to 8 tapes each holding 20GB without compression), bought for a song on eBay, with BackAgain/2000, which handles EAs, open files, and networking security information just fine (now abandonware, but demo versions all over the place and registration keys available without looking too hard). The last lot of brand-new IBM 20GB tapes cost $3.50 each. PCI-bus SCSI cards are cheap.

Now time has changed. The standard user doesn't seem to have SCSI tape drives in most cases and people are shifting to use CD/DVD or just USB/MSD or simple "hard disk" drives for backup. This makes it easy to come up with even a rexx-based solution. Indeed, it is something I am interested in a) using and b) perhaps developing.

Hard disks are too fragile for reliable backup -- no more reliable than the "working" media. Typically they are connected to the same controller as the "working" media, so the backup could be as corrupted as the original. (My tape drive and backup software are in a separate machine on the network, so I don't have the potential problem of a controller failure corrupting both backup and original.) External hard disks usually are not connected to the same controller, but they are subject to mechanical shocks that can render them and the data they hold useless; don't ask me how I know.

Even double-sided DVDs have a capacity too low for serious backup of large amounts of data.

If you really want to use hard disks (either internal or external) for backup, I suggest using RAR. And you could still burn the resulting RAR archives to DVD.

<snip>