Hello, we are runnig a Navision 3.70a Server with a native database consisting of 6 parts, each sized 10 GB. DBMS cache is 40000 kb, commit cache yes, object cahce 20000 kb, the maximum number of users is 52. The server hardware is a Intel Xeon 3.0 with 4 GB ram and W2K3-Server. Our problem: The system is running properly unless the last user is logged out. At that moment each slave.exe is releasing its memory one after the other. This can last up to half an hour. During this time a user trying to connect is getting an error message. Does anybody face the same behavoir of the Navision database? Or does know how to avoid this problem? Looing for any help or hint, thank you Marcus Weichelt
Follow this thread for more info: http://www.mbsonline.org/forum/topic.asp?TOPIC_ID=9866
Is this problem specific to 3.xx? We have a similar configuration with 2.60 and have never experienced any such problems (aside from the usual “…restoring list of free blocks” after restart, which takes but a few seconds).
Hi Heinz. I already experienced the same when using 2.60 and 2.65 so it shouldn’t be a version issue. And it was the same server machine throughout these versions.
Hello Nelson & Heinz, we are already working with the “one client always stays connected method”, but on the long run this is not satisfying. It´s only a workaround. You end up trying to fix up symptoms not solving the problem itself. The problem seems to be the management of the navision database/server cache. What I actually was looking for is a solution to avoid this behavior without workarounds like alltime open sessions. cu, Marcus Weichelt