Re: Design question
Tech-Archive recommends: Fix windows errors by optimizing your registry
Date: Wed, 25 Feb 2004 18:50:55 -0500
Yes... that's generally what record locking was invented for. You need to
obtain an exclusive lock on the resource, in this case a record in a table.
An alternative that comes to mind is that you can have a service running
which fetches records for processing. Once a record was fetched and handed
to one of your machines it (the service) wouldn't hand it out again to the
next machine that requested one.
Tom Leylan
"Ruslan Shlain" <rshlain@hotmail.com> wrote in message
news:OZJC1L%23%23DHA.3668@TK2MSFTNGP09.phx.gbl...
>
> Hello Everyone,
> Looking for a better way to do this.
>
> I have several separate processes running on 5 + machines on the network.
> These processes are identical. What they do is, they look at a table in
the
> DB and based on entries they perform some tasks.Once entry is read it
> changes a status. My problem consits of the following. If 2 processes read
> an entry from the table at the same time and one of them changes it status
> and then the other comes in and changes same entry they both will be
working
> on the same task. Is there a way that i can force "something ", so that
> above scenerio happens 2 of my processes do not work on the same entry.
> --
> Ruslan Shlain
> nAlliance Corporation
> www.nAlliance.com
Relevant Pages
- Re: Mass delete files and directories
... The fact is that shifting those twelve bytes upwards by two bytes is *always* the correct thing to do, whether your existing code is currently giving problems on a specific machine or not, because it causes the code to work correctly on all machines. ... In fact it is absolutely essential on all machines if you want to use the lpszProgressTitle entry or perhaps the hNameMappings entry. ... It is of course easily possible to test in code for such an eventually (the LongWord alignment change, not the lottery ticket or the lightning!) but I think that doing so would be going way over the top and I reckon that simply ensuring that you *always* shift the appriate block of data up by two bytes will be fine. ... GP hasn't said whether any of the three "working machines" are SP2 or not, but I suspect that SP2 is the problem here, possibly something to do with enhanced protected files machanism or something else that is causing either of the two Longs following fFlags to be populated with something other than zeros at some point in the proceedings, which then causes the problem because those data bytes are shifted down by two bytes from where SHFileOperation expects them to be. ...
(microsoft.public.vb.general.discussion) - Re: Constructive Feedback on things to make PAPA even better
... to respond to the below as I felt the same way after my first PAPA ... machines, ... play a round and was able to schedule a start time for each entry. ...
(rec.games.pinball) - [HPADM] Re: sdsadmin ?
... implies that we can use this to create an /etc/disktab entry for the dis we ... machines don't have a man page for this command. ...
(HP-UX-Admin) - Re: Mass delete files and directories
... The fact is that shifting those twelve bytes upwards by two bytes is *always* the correct thing to do, whether your existing code is currently giving problems on a specific machine or not, because it causes the code to work correctly on all machines. ... In fact it is absolutely essential on all machines if you want to use the lpszProgressTitle entry or perhaps the hNameMappings entry. ... It is of course easily possible to test in code for such an eventually (the LongWord alignment change, not the lottery ticket or the lightning!) but I think that doing so would be going way over the top and I reckon that simply ensuring that you *always* shift the appriate block of data up by two bytes will be fine. ... GP hasn't said whether any of the three "working machines" are SP2 or not, but I suspect that SP2 is the problem here, possibly something to do with enhanced protected files machanism or something else that is causing either of the two Longs following fFlags to be populated with something other than zeros at some point in the proceedings, which then causes the problem because those data bytes are shifted down by two bytes from where SHFileOperation expects them to be. ...
(microsoft.public.vb.general.discussion) - Re: List of Everyones Pins
... to a database that I set up that will allow any collector/player, ... entry, or the email thing is optional. ... send easy commands to list all the entrys by users, machines, ...
(rec.games.pinball) |
|
