Windows 10 Caching while writing to linux server issue
The issue I am having is we have a piece of Point of Sale Software that runs from a Linux server, the software is launched from windows workstations, when we create a transaction on a Windows 10 Pro Machine and the software asks to print the ticket (this happens randomly) it sometimes prints a blank ticket. I have spoken to the programmer on how the data is written to the database and he said it is possible that windows 10 is just not writing the data to the database fast enough so the software is printing what ever is in that field of the database which is still blank. This does not happen on the Windows 7 Machines on the network. So he said if it is some type of Caching issue. If someone could point me in a direction to test a few things I would greatly appreciate it.
Can you run the application POS application in compatibly mode ?
ASKER
Zaheer, No I cannot I thought about that it does not work.
David, The issue is not that printing is intermittent, the issue is that for some reason windows is not writing to the database on the server quick enough and the software goes to print and the field in the database is still empty. I don't think it is a networking issue per say, I think it is a windows 10 Issue but I am not sure what to look for specifically. Again from a Windows 7 Machine there are no issues, and this all started after replacing 2 windows 7 machines with windows 10.
David, The issue is not that printing is intermittent, the issue is that for some reason windows is not writing to the database on the server quick enough and the software goes to print and the field in the database is still empty. I don't think it is a networking issue per say, I think it is a windows 10 Issue but I am not sure what to look for specifically. Again from a Windows 7 Machine there are no issues, and this all started after replacing 2 windows 7 machines with windows 10.
Still a code bug.
The code design you just described is full of race conditions, so will be dependent on many factors.
Simple code fix is to add some flag to your print data, which gets set to say data writing to database is complete.
Or better, just use correct row locking logic at both producer + consumer ends will also work.
The code design you just described is full of race conditions, so will be dependent on many factors.
Simple code fix is to add some flag to your print data, which gets set to say data writing to database is complete.
Or better, just use correct row locking logic at both producer + consumer ends will also work.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Likely best to run tshark on your Linux machine (as tshark is a snap to setup under Linux) to trace packet flow + track payload of your print jobs to see where data is lost.
This will be a long + complex process.
If you're new to this type of work, great if you have the time/money/will to learn the process yourself.
Is time to fix problem must be quick, hire someone to do this work.