Jim, SP5 and KR2 for AX 3.0 solved all data synchronization problems that we were experiencing between the AOS's (when we were on Ax3.0 SP4 in a 3 Tier Environment with multiple AOS's). I highly recommend downloading these 2 upgrades. The download is an executable (.exe) but it is a self extracting zip file. By clicking on the executable you are not running the update. You are extracting the update and all of the associated information (such as the "Fix List" word document, The Read Me, and the installation checklist). KR3 is now out, so you might be better off going to KR3 after installing SP5.
After downloading them, extract them to your hard drive and read all of the information that is extracted. If I remember correctly, these documents contained a wealth of information. To give you an idea of how extensive this fixlist is, the SP5 Fixlist alone is 185 pages in Microsoft Word. Best Regards, Mike ----- Original Message ----- From: jimprendergast2006 To: [email protected] Sent: Wednesday, November 15, 2006 12:49 PM Subject: [development-axapta] Data synchronization in AOS Cluster There has been some discussion around this topic before in this Group (I've paraphrased below). Any comments or suggestions are welcome. Here's the scenario: We're running Axapta 3.0 with SP4 in a 3-tier environment, I'm guessing that the problem is with Cache synchronization between our two AOS servers configured in a cluster. The AOS's are on separate servers (let's call them "A" and "B"), with a thirdmachine that runs SQL server. The application folder is placed on a shared directory of "A" and has no access restrictions on it. All the client configurations are set to use the cluster and the load balancing works properly (the client sessions are instanced alternatively on "A" and "B"). The wrong thing is that sometimes the data on "A" and "B" are not "synchronized" (i.e. records inserted by one user working on "A" are not seen by other users working on "B"). Everything seems to solve after an AOS restart, but this is not a practical solution since the problem may occur many times in a day. Any solution or explanation? Is this issue solved in another version or release? Is there a caching feature to disable/change? ... Table caching level on tables where I experience the problem ("none", "found", "notInTTS", "found and empty", "entire table") Whole installation caching parameter "-internal = maxCacheLevel:"? Should we consider some deletion policy for the .aoc files? Should we disable "data preload"? etc ... ------------- A solution that I've seen proposed is: You can clear the caches without having to change caching options or restart your AOSs. Turning off caching is really quite a drastic and severe solution, and I wouldn't recommend it at all. Look for the sysFlushAOD and sysFlushData action menu items. Create new versions which run on the server (RunOn property set to server). If you are logged on to a particular AOS, and run those two, then it will flush the application and data caches respectively. You will still need to restart each affected client (or run the same two menu items client-side for every client) as they also cache application objects and data. Thanks to: Andrew Jones - HGH Business Consultancy Unfortunately, requires intervention after obsolete data has already been encountered and the user will never know until its too late. ------------- Jim Prendergast [Non-text portions of this message have been removed]
