[CXX-496] client & integration tasks time out after a successful run on windows Created: 27/Jan/15 Updated: 28/Jan/15 Resolved: 27/Jan/15 |
|
| Status: | Closed |
| Project: | C++ Driver |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | legacy-1.0.0 |
| Type: | Bug | Priority: | Blocker - P1 |
| Reporter: | Adam Midvidy | Assignee: | Adam Midvidy |
| Resolution: | Done | Votes: | 0 |
| Labels: | legacy-cxx | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Description |
|
See the abrupt time jump here: https://mci.10gen.com/task_log_raw/mongodb_cpp_driver_windows_64_msvc2010_dbg_dyn_dyn_2d738d6a1064920151f380513363027228258caf_15_01_23_23_38_04_integration_test_latest_windows_64_msvc2010_dbg_dyn_dyn/0?type=T#L843 Seems to be related to a failed attempt to shut down MO. I have a patch here: (https://mci.10gen.com/version/54c666ab3ff122791300036a_0) that just relies on the killall_mci expansion to pwn MO, since it kills all python processes and any mongo(s|d) that MO may have started. However, I think I want to investigate a bit more on a spawnhost first as giving up on clean shutdown is a bit hacky. |
| Comments |
| Comment by Adam Midvidy [ 27/Jan/15 ] |
|
This appears to have been fixed by removing an erroneous 'cd' command from the trap handler on windows when we start mongo-orchestration. I'm not exactly sure why that fixes the issue - but it does, so I am going to close this out. |