Hello Jeff, I am thinking of trying something like what you were testing back in 2014 so I'd love to hear how it has gone the last couple of years? Thanks, Alan
On Sun, Nov 2, 2014 at 9:01 PM, Jeffrey Kain <jeffrey.k...@gmail.com> wrote: > I’m doing some experiments with running network-intensive client > applications alongside 4D Server as opposed to running them traditionally > on a dedicated client on the LAN. My performance results are unequivocal - > I’m seeing about a 2.5x performance increase when running client alongside > server, as opposed to running client on the gigabit Ethernet LAN. > > In addition to the raw performance improvement of these dedicated client > applications, the entire system benefits since the data required by these > clients doesn’t saturate the network. > > Sounds great, right? There are two troubling observations, however. Since > moving to this configuration, I’ve noticed: > > 1. Client memory usage on these workstations running next to the server > increases linearly over time. By linearly, I mean that if I plot memory > usage over time I get a perfectly straight line. I find it hard to believe > that the application uses memory like this given the eb and flow of how the > system is used (peak times, dead times, etc.). It looks like a memory leak, > but I hate to throw that term around. Eventually I have to restart these > clients when they get close to the 2GB limit. > > 2. Server memory usage increases much faster when I have a few of these > local clients running than when I don’t. The server has something like 80GB > of cache memory allocated, and prior to testing this configuration it would > take well over 30-45 to fill the cache. Now it happens in just 10-12 days. > > 3. The server has hung twice in the last 6 weeks (i.e. stopped processing > data but not crashing per se). Prior to this configuration, the server was > very stable. > > I realize that correlation does not imply causation, but I’m wondering if > anyone else has experimented with this kind of configuration and if you’ve > observed anything similar? > > Thanks! > > Jeff > > > P.S. - The environment is as follows: > > Server: Windows Server Enterprise 2008 R2, 192 GB RAM, 1 TB PCIe SSD RAID > array > 4D Server: 13.5 (also tested 13.4), 64-bit, 78GB allocated to cache, ~110 > users/900-1100 processes > Data: 90GB, Index 18GB > Clients: All Windows 7 > v > — > Jeffrey Kain > Senior Laboratory System Architect > Health Diagnostic Laboratory, Inc. > > > ********************************************************************** > See how easy it is to extend your 4D solutions to Web and mobile. New > opportunities await you with 4D v14! > > 4D Internet Users Group (4D iNUG) > FAQ: http://lists.4d.com/faqnug.html > Archive: http://lists.4d.com/archives.html > Options: https://lists.4d.com/mailman/options/4d_tech > Unsub: mailto:4d_tech-unsubscr...@lists.4d.com > ********************************************************************** > ********************************************************************** 4D Internet Users Group (4D iNUG) FAQ: http://lists.4d.com/faqnug.html Archive: http://lists.4d.com/archives.html Options: http://lists.4d.com/mailman/options/4d_tech Unsub: mailto:4d_tech-unsubscr...@lists.4d.com **********************************************************************
