Helios10 to Helios11 upgrade problems

[dlparker]

My apologies for the length of this post, but this involves some troublesome issues, not all of them technical. Here goes:

I'm an independent computer tech (over 26 years IT industry experience, including mainframes, PCs from MS-DOS 4.2 onward, UNIX SVR3/SVR4, and Linux from kernel release .9x through current) with a client who has two salons, both running Helios software on Windows 2000 PRO machines. I took over the account after the stores were up and running, and over the course of about two years have replaced all the machines (starting with the eMachine running Windows ME, which died a typically ugly death, taking all the data with it and leaving only an 8 month old backup) with ones running as close to identical hardware as I could get - not bleeding edge, but reliable machines built by my distributor as the need arose - for business use, 512MB memory, Windows 2000 Pro. Both stores recently (a/o May 2-3) upgraded from 10.2 to 11.4. I was supposed to perform both upgrades, but an employee at one of the stores was instructed by another employee to perform the upgrade, so she did as she was told. I got called when the system wouldn't work, so I went onsite a day ahead of schedule and got things straightened out with not much hassle. This store has 3 workstations, and the transaction DB (transact.mdb) is about 1.47GB. Helios11 has been running fine at this location since May 2. They are by far the busier of the two stores.

The next day I performed the upgrade at the other store - 2 workstations, transact.mdb around 600MB before the 10.2 to 11.4 conversion and about 450MB after the conversion. This one has been nothing but trouble. The upgrade process itself took over 6 hours, mostly running comparedb and the MS ODBC repair utility and doing two 10.2 to 11.4 DB conversions. All per the instructions of Helios tech support. The most common symptom (and I say "symptom" to emphasize the extent of the wiggle room in interpreting the symptoms and diagnosing the problem) was the error message "Unrecognized file format on transact.mdb", sometimes on startup, sometimes right in the middle of entering a transaction (I was usually not on site for the latter). The error number was always -2147206461, which looks to me to be about the maximum negative value an integer field can hold. Been a long time since I've studied that stuff. At any rate, totally meaningless. The helios tech support solution is roughly as follows:

Run comparedb. Select all DBs, check all 4 boxes.

Run the MS ODBC repair utility.

Kill the MS ODBC repair utility (or sometimes comparedb) when it hangs (15 minutes or so with no disk activity and 90-99% CPU utilization by the utility in use).

Reset the permissions on helios11 and all subdirectories to read/write/nuke for everyone.

If it happens again, do the 10.2 to 11.4 DB conversion again.

Start the process over when it happens again.

At some point we got it working. And it did work - for about eight days or so, May 4-12. When it started doing it again the client said, "Go back to 10.2". I did this on Friday morning, May 12. Fortunately they had not yet sent back the 10.2 parallel hardware keys. The store manager decided on Monday, however, that reentering the one week+ of lost data from Helios 11 was a more unpleasant prospect than risking further v11.2 problems, so they asked me to switch them back. At this point they've been running since last Monday, May 12, with no problems.

The solution to this problem, according to Helios tech support, is to archive the data. The owner absolutely does NOT WANT TO DO THIS. As a business owner, that is her perogative. And it seems to me a sensible business decision to want to have all the customer data available to the application. To archive THIS transact.mdb (which is about 1/3 the size of the transact.mdb at the other store which has had no problems) in order to fix a problem, the cause of which no one can pinpoint, seems somewhat arbitrary to me. I don't believe that a software vendor should be dictating business decisions based on technical limitations of their software.

I've got extensive experience on other platforms, where when something goes wrong, there's usually a log file that will tell you what happened, or error messages that mean something, or a database tools that can give you a little information. I have no experience with MS ODBC databases or tools, but this looks to me like an ODBC problem, or at least a problem involving corrupted data. Not being able to do anything other than watch it break, however, I have no way of knowing.

THE QUESTION: Has anyone had a similar experience, or does anyone know how one might determine what is actually causing the problem?

	Buy Spray Tan Solution www.spraytan.com - Wholesale spray tan solution for tannings salons and mobile spray tan professionals. Lowest prices on premium Bronzz™ spray tan solution. Booth or airbrush tanning.

[caribbean tan man]

what #&**&^ are u seriously trying to ask a question in that mumbo jumbo? all i remember from it is 6 hrs to upgrade! what r u using a 52k dial up.

im no computer wiz and find it very simple to upgrade and did.

you can save the data folder to the desktop uninstall helios and then reinstall the newest version then copy/paste the dat folder into the new version............DONE.