Something is eating entropy in test environment
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mir |
New
|
Undecided
|
Unassigned |
Bug Description
10:34:58 13: [ RUN ] TestClientInput
10:34:58 13: [2016-03-17 10:34:58.778663] mirserver: Starting
10:34:58 13: [2016-03-17 10:34:58.780193] mirserver: Selected driver: dummy (version 0.21.0)
10:35:28 13: ERROR: ../../.
10:35:28 13: Dynamic exception type: boost::
10:35:28 13: std::exception:
10:35:28 13:
10:35:28 13: unknown file: Failure
10:35:28 13: C++ exception with description "Failed to start server thread" thrown in SetUp().
10:35:28 13: [2016-03-17 10:35:28.880620] mirserver: Stopping
10:35:28 13: [ FAILED ] TestClientInput
10:35:28 13: [ RUN ] TestClientInput
10:35:28 13: [2016-03-17 10:35:28.897809] mirserver: Starting
10:35:28 13: [2016-03-17 10:35:28.898771] mirserver: Selected driver: dummy (version 0.21.0)
10:35:58 13: ERROR: ../../.
10:35:58 13: Dynamic exception type: boost::
10:35:58 13: std::exception:
10:35:58 13:
10:35:58 13: unknown file: Failure
10:35:58 13: C++ exception with description "Failed to start server thread" thrown in SetUp().
10:35:58 13: [2016-03-17 10:35:58.939169] mirserver: Stopping
10:35:58 13: [ FAILED ] TestClientInput
Looks like lp:~vanvugt/mir/run-without-entropy running on the same m/c kills us.
Correct. This problem is expected (and I deal with it on my desktop every day too).
I am conscious of this problem and only ask CI to update the branch when necessary.
The solution is really just to land the branch, so all branches then have the fix. This is a duplicate of bug 1541188.