Workflow Soluiton

I was running Workflow Designer 7.0 on my machine (off-server) with no problems.  It was a peppy interface and worked great.  When I upgraded to 7.1 on a new server, I also upgraded the software on my machine.  I'm now noticing significant sluggishness while designing workflows.  Sometimes the program hangs for a few seconds while I'm trying to move components, debugging performance is a lot slower, and sometimes the program locks up when I'm searching for components via the search box, usually after I have a lot of time consuming unsaved work (grr!).

I'm running Windows 7 x64 w/ 8GB of memory on a Core 2 Duo so it shouldn't be anything on the hardware end.

Anyone else see similar issues? I'm wondering if I need to do a complete uninstall/reinstall of Workflow Manager, or worse, an OS reload.

Is there any other software on the server (database, SMP, etc.)? That'd definitely affect overall performance.

Is LogicBase.ToolCore (project designer) the definitve resource hog on the computer?

I have a Windows 7 64 bit machine with 12 gig of memory and haven't noticed any lagging issues.

In both 7.0 and 7.1 I run SQL on the same server, but it has 8 cores and 56GB of memory so that's not the issue.  The workflows also run great when published, it's just when I'm building/debugging on my machine where it's ugly.  I'm wondering if the upgrade did something wonky.

Dfrancis, your experience isn't typical from what I've seen.  On my VM dev box (much lower specs than yours) running WF 7.1 GA , most projects open immediately.  I have one very large project (50+ models) that takes longer to open, and can be sluggish when opening a model, but once the model is open, performance returns to full speed.  Certainly nothing like a lockup or hanging when moving components.

You should be aware that there's a new translation service added to the product.  During debug (and publication), the workflow server will attempt to contact the translation web service on the default SMP defined in Workflow Credential Manager while rendering each web page.  If that SMP server is down, it can add ~15 seconds to rendering each page, as it must wait for a time out to occur.  You can see this error in the Workflow Log viewer.  I haven't had time to learn this new capability, so I haven't developed any best practices for configuring my dev environment.  At the moment, I either make sure I have a valid SMP target, or rename the translation plugin.dll found in the workflow designer/plugin folder.

Do you observe any sustained CPU spikes when this problem occurs?  If so, what's the top process at the the time?  Typically, when I open a very large project, there's a lot of C# compiling that has to occur (csc.exe).

Less likely, but worth checking, see if your SQL server is suddenly demanding much more memory than it did in the past.  It's not possible to see all of SQL's memory consumption in process manager for the mssql process, but if you look at the Performance tab > Memory graph, the total memory consumed will include the additional memory reserved for SQL data caching.  If you stop SQL server and that amount goes down by more than 4GB then you know where to troubleshoot.  I posted some articles on troubleshooting SQL performance with Perfmon many years ago, they should still be in the KB.

If none of the above are applicable, there's also the chance that you have a bad spot on your hard drive.  The OS will try hundreds of times to read a weak spot and won't tell you that there's a problem if it's finally successful.  The symptoms are typically freezing for a few seconds with very little I/O load. The quick test/fix is to run scandisk with the repair option enabled and reboot when prompted.