This test program set is intended to demonstrate a problem with NOVELL. Under some circumstances NOVELL will return incorrect data on a read. One workstation writing information immediatly followed by another workstation doing a read of the same information will sometimes result in the reading workstation getting back part of a new record mixed with part of the old record. TESTW will write a pattern. The pattern alternates between writing a record full of 1's then 2's then 3's etc. You should always have a record containing all the same character. Use the ALT key to escape out of either program. Run the WRITE program on one workstation, run the READ program on one or more workstations. Start the WRITE program first. If you are going to see a error you should expect it to happen somewhere between 100,000 and 500,000 reads. The problem shows up more on a heavily loaded server and may not show up if this is the only job reading/writing on the server even if your server is one with the problem. Note that there are some known bugs that can cause this problem under specific shell setups (workaround is the CACHE BUFFERS=0 statment) and there is a bug with Server CACHE that can cause this type of problem under Netware 3.10a. Novell says that there are no known bugs that can cause this problem under the current versions of netware running the latest shells with the latest drivers. I have been able to reproduce this on several servers running different drivers, shells, etc. Source code included and should be able to be compiled under TP6 or TP7. If you run this program set, please E-Mail 70007,3427 on CompuServe and report your results. Anyone who E-Mails their results to me will be put on my mailing list and will be kept updated as we find more info about this problem. The information I need is: Did the test set fail on your server? What is the date/time/size of your SERVER.EXE file (from the boot drive) What is returned when you type CONFIG on the server console. What is returned when you type MODULES on the server console.