Closed kzemek closed 8 years ago
Indeed. I've pushed a fix, which also has a test for the 2-node case. I verified that it failed before, and it seems to pass with the last commit - see PR #26
Seems to work for me. After updating to locks
from the new branch I had one case where both leaders surrendered after the first connection (no netsplit in the meantime); I recompiled immediately and couldn't reproduce it after.
Ok, I've merged the PR. Thanks for your help!
I found an unexpected behaviour of locks_leaders; I confirmed it using test_cb.erl. The repro is very simple:
test_cb
on nodes A and B; each process is a leadertest_cb
processes will become the leadertest_cb node A logs test_cb node B logs