Results 1 to 7 of 7
  1. #1
    Join Date
    Aug 2006
    Posts
    2

    Question Unanswered: DBase and clipper

    Dear all,
    I recently joined a retail firm running a dbase data base (centralized) and a Point of sales system built with clipper. We have 25 POS machines and 10 desktops.the problem is that the whole system keeps going down. we have already identified that the server and network is ok. IS the system too old to handle the work load of these 25 POS machines? I know little about clipper and dbase so please help me out.
    And i ask the administrator to forward my question to any relevant forums.

    Thanking you,
    chris

  2. #2
    Join Date
    Feb 2004
    Location
    In front of the computer
    Posts
    15,579
    Provided Answers: 54
    I missed this message somehow...

    Too much depends on your archtecture to be able to give you a simple answer. Can you give us a bit more detail on how things are set up? I'm curious as to whether the POS terminals just terminals with a cable to a central computer, or if they are computers with a cash drawer attached. That will make a world of difference.

    What exactly do you mean by "going down" ? Do all of the systems simultaneously spit blue smoke and go cold? Does one POS station die an hour? Do you get bad data? There are lots of choices here.

    -PatP

  3. #3
    Join Date
    May 2005
    Location
    CHicago
    Posts
    26

    DBase and Clipper

    [QUOTE=chriswhereyou]

    When Clipper reads rords from a database it copies all the records to the local drive. There it swaps them to a temp file (SET TEMP=TMP ot SET TEMP=TEMP) many clipper linkers will let you burn a location but the location needs to exist. Next all those records can bog down your network. Advantage clientserver system will send only the records called for. Thus reducing the network trafic,
    There are other reasons Clipper would stop. Append does not wait for a record lock it just checks for one. The use of resevered words can cause problems. A bad sequence of lock / unlock can do it. In other words only an experiance Clipper programmer, reading the code, can see all of the reasons.

  4. #4
    Join Date
    Feb 2004
    Location
    In front of the computer
    Posts
    15,579
    Provided Answers: 54
    Actually, Clipper only copies the rows that it actually needs to read or write to the local PC. For poorly written code, this can mean the entire file (often for every command), but that isn't always the case because some Clipper programmers wrote pretty decent code.

    You are correct that it Advantage will help some applications that are written to use Advantage to run better. Advantage by itself won't do anything for an existing program that doesn't know how to use it.

    Most Clipper programs can be decompiled if necessary. Nearly all Clipper programs can be tuned to perform better by someone that understands them, even without access to the source code.

    You are definitely correct in that there are many things that can cause performance problems for a Clipper application. I'm trying to get enough information about the configuration that the original poster is using to be able to offer some good suggestions. One of those suggestions might well be to get a programmer involved. Another might be to replace the whole system outright. If possible, I'd much rather find a simpler answer that gets 80-90% of the benefit for 10-20% of the cost (if I can find such an answer).

    -PatP

  5. #5
    Join Date
    Aug 2006
    Posts
    2

    Details on the POS system

    Dear All,
    Thank you for the replies. I give below some details about the system in question. the pos terminals are connected via a novell server that acts as a domain controler by having users that are authorized. This server also holds the main database that the POS program accesses. the dbase database holds about 65000 item records. and yes the pos terminals do have customer displays and cash drawers too. so the print goes both to the pos printer and the customer display.
    The way the system goes down is that when the user enters product codes via the barcode reader the systems cursor keeps blinking and this travels throughout the pos machines until none are accesible. this happens about once a week. then wat i do is to cancel all the connections from the novell server and reorganize the system. wat the re organizing does is it gats all the dbase files in order. the re organizing is usually done in the morning before the shop opens. we used a network meter to check for collissions and errors but they look normal. could it be that the system cant handle 65000 recs?

    And please do tell me how to optimize the clipper software without the source code. i think i can manage with my SW development knowledge. The mamagement had tried to get a new system but after 2 years they had to give up the project coz the vendor was unable to meet the quality standards required. so i am now stuck with the exizting one till the new system is develpoed which could take another year.
    If u need any other info please do ask

    thanx in advance
    chris




    Quote Originally Posted by chriswhereyou
    Dear all,
    I recently joined a retail firm running a dbase data base (centralized) and a Point of sales system built with clipper. We have 25 POS machines and 10 desktops.the problem is that the whole system keeps going down. we have already identified that the server and network is ok. IS the system too old to handle the work load of these 25 POS machines? I know little about clipper and dbase so please help me out.
    And i ask the administrator to forward my question to any relevant forums.

    Thanking you,
    chris

  6. #6
    Join Date
    Feb 2005
    Posts
    14
    There is nothing in what you have noted that would point to the CLIPPER program as the cause, it may be, but that can't be deduced from what you have stated. I have a fairly large CLIPPER app and we used to be on Novell and it ran fine, the only time I had problems was when the OS got updated on the desktops. We did constantly recreate indexes. How offen do you bring the server down and run maintenance. If the app has been running cleanly for a number of years, what changed to cause this issue. 6500 recs is nothing compared to the volume of data that I push around so the rec count isn't an issue. The part about reorging your system is curious, if it is written properly you shouldn't have to reorg dbase tables? You may be having an INDEX issue, and the reorg is getting your indexes clean. NTX indexes are notoriously flaky with larger amounts of data or volatile files. You may need to recreate your indexes daily? Without looking at the source code where all just guessing, nornally I would trace your issue and see where it gets lost.

    Jim C.

  7. #7
    Join Date
    May 2005
    Location
    CHicago
    Posts
    26
    Clipper can handle 1 million indexed records and 2 geg of unindexed data.
    Link in the debug object and see what is happening.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •