Plateforme Level Extreme
Abonnement
Profil corporatif
Produits & Services
Support
Légal
English
Appended record missing in another data session
Message
De
30/06/2004 21:51:04
Neil Mc Donald
Cencom Systems P/L
The Sun, Australie
 
 
À
30/06/2004 08:50:09
Walter Meester
HoogkarspelPays-Bas
Information générale
Forum:
Visual FoxPro
Catégorie:
Problèmes
Divers
Thread ID:
00902364
Message ID:
00919471
Vues:
40
Hi Walter,
>I don't share that experience. I've got quite a few applications out there that run various kinds of networks and did not have a single identified report of corruption within 5 or 6 years. Now I am not the system administrator and can't see how the server is configured but I think it is safe to bet at least a great portion of them did not follow those reccommendations.

Our main application is very complex Billing system that pushes VFP to it's limits.

Your comments sound exactly the same as the contractor hired by one of our recent new clients (he had more MS Certifications than anyone I have ever seen), his comment was "These OS setup requirements are a load of s**t", and he didn,t do the required disabling of the caching etc, the system ran like a dog, several crashes per day.

I instructed our client to instruct their contractor to apply the required settings as specified or don't call us expecting free support, they did so.
The system has now run of 7 weeks without a glitch, the support calls now are only of the "How to variety".

The system admins you deal with MUST know their stuff, that is why you aren't experiencing any problems or your apps don't stress the systems.

Another problem we experienced in the last 6 months was where an outside contractor had set the system up (he also didn't apply the cache disabling etc), but hadn't told the client the admin password, probably so as to milk more fees from the client.

The client experience a blackout, but they had a good UPS which gave them abt 1 hr on batteries, the girls wanted to go home on time so they completed the end of day postings (about 400 invoices & 600 payments, a small office).

They knew they had to shut the server down, but without admin rights they didn,t have a shutdown option, so in their wisdom they just logged out then turned the server off.

Write caching was enabled on the server Raid, also the primary raid drive had failed and they hadn't bothered to tell anyone, this resulted in major corruption of the databases. If the write cache was disabled the updates probably would have been written to disk i.e. no corruption.

As a footnote, a properly setup network is a rarity these days, as there is always some cowboy out there who will cut every corner in the book to win the job, so this combined with the actions of the operators is what has lead me to my current stance re: caching & oplocking.

I hope this helps you in the future.
Regards N Mc Donald
Précédent
Suivant
Répondre
Fil
Voir

Click here to load this message in the networking platform