At Thu, 8 Aug 2013 19:41:33 -0700, Kai Zhang wrote: > > v2: > - add a patch to fixed wrong return code of zk_join > - reduce the interval of getting lock rety from 1s to 50ms > - fixed wrong calculation of queue_pos > > Kai Zhang (2): > zookeeper: zk_join() should return 0 on success and -1 on error > zookeeper: remove master from zookeeper > > sheep/cluster/zookeeper.c | 241 ++++++++++----------------------------------- > 1 file changed, 54 insertions(+), 187 deletions(-) I wonder if the second patch should go into the 0.7.0 release. The cleanup looks not trivial and, IMO, needs more tests. What do you think of it? If you have a enough time to test the zk driver before the 0.7.0 release, I'm happy to include the patch into the release. Thanks, Kazutaka |