© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
0e1f616
(Get the code!)
I've been chasing this off and on for a few days now. Here is what I know so far
1. there were some races, i've fixed those
2. this absolutely happens with the packaged mongo, and absolutely does not happen with our tarball
3. it is not actually a problem with TLS, i've added debugging to crypto/tls and we are negotiating the same ciphers
4. i believe it is either
a. a race between cleaning up the mongo server and the cleanup actions themselves that try to remove the juju and presence tables. once they fail, then the next test does not run cleanly, and more failures propagate from that.
b. a packet we send to mongo which isn't valid for 2.2.4, causing it to slam the connection closed.
5. if I run the tests under the race detector, they pass, which gives me more ammunition to think it is a race on the cleanup actions of the test.