Hi everyone,

We're running GW 8 SP2 on NW 6.5 SP8 and NDS build 115 (the latest release version.)

Every once in a while, we encounter a problem where the POA will start to run slowly and things won't sync well between the Datasync server and the phones (currently have ~15 devices (iPhones and Droids.)

Whenever that happens, we'll notice that on the SOAP monitor screens on the POA, there will be a number of SOAP threads stuck at "Handle Request" state for 500 seconds plus (up into the multiple thousands of seconds.) When you view the details of those threads, they'll usually be in a "Current Transaction" of "getItemsRequest" or "CursorCreate". The threads are in use by the DataSync server for the same three users each time (users with larger mailboxes,) and there are multiple threads open for these users in this state.

When this happens, it slows down emails for everyone on that POA because the number of CS threads in use goes really high (usually 10-20 in use.)

Before too long, clicking on the thread links in the SOAP monitor screen will result in "Invalid URL" error messages, and soon after that, all 80 of the SOAP threads will lock open and nothing will sync over SOAP. We also can't restart the POA because it won't shut down (keeps cycling messages about waiting for the SOAP threads to terminate, but they never do,) so we're forced to restart the entire server (which houses 6 POAs for around 1500 users.)

The option of "delete the user from NDS" isn't a viable one (for one, we've tried it before and it didn't help, and for two, that's just not a viable solution for a production system.)

The option of "restart the POA" isn't a viable solution either for the same reasons, as it makes email unavailable for all the users on the different POAs since the POA won't close gracefully and we have to restart the server (which only serves to increase the rumblings of the users requesting to not use GW anymore.)

Any ideas? This really isn't an acceptable problem to be having.