I am trying to track down the cause of an AcrMap application freeze. I am able to replicate the issue using a very complex mxd with mxdperfstat93 with default scale parameters. We were targeting the OS because the tests had no issues on Windows 2008 sp2 where it was failing on Windows 2008R2. Unfortunately, the OS was not the only variable. The 2008 sp2 machine is situated right next to the database (1 hop with a tracert). The 2008R2 machine is connected to the database with 1000 Mbps fiber line and has 4 hops with using tracert to the database. To test the OS being the cause, we built a VM with Windows 2008 sp2 and placed it in the same data center as the 2008R2 machine. We expected the mxdperfstat test to succeed but it failed the same as the 2008R2 machine. Because, we are not testing apples to apples (VM vs a physical server and different memory/cpu allocations) we do not want to jump to the conclusion that it is the network.
I am doing client-side sdeintercept logging using direct connect. There are 2 logs that are output during the operation. Below is a snippet from each one before failure. Included is the end of the logs. The log ends abruptly without any error message.
Error generated in DOS window by mxdperfstat: buffer_alloc stream buffer block: errno = 8
There are 3 mxds that I test. The 2nd mxd has scale dependencies on some of the major feature classes and fails in at a different spot but with the same errors. The third has scale dependencies set for all major feature classes (1:50,000) and succeeds in all tests on any platform tested.
Research on both the mxdperfstat and sdeintercept errors have yielded nothing (I have not found the descriptions for these errors). I don't want to make assumptions about these errors. Any insight would be appreciated. Thanks