Github user revans2 commented on the issue:
https://github.com/apache/bookkeeper/pull/116
@eolivelli I get them too. It may be differences in the default value of
ulimit. Probably the simplest thing to do it to try and trace down who is
leaking all of the file descriptors. It most
Github user revans2 commented on the issue:
https://github.com/apache/bookkeeper/pull/110
+1 still looks great.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and
Github user revans2 commented on the issue:
https://github.com/apache/bookkeeper/pull/110
@eolivelli Validation of the client side principal in ZK is done through
ACLs. The simplest thing to do in our case is to have a white list of clients
(possibly with regular expressions). If
Github user revans2 commented on the issue:
https://github.com/apache/bookkeeper/pull/110
@eolivelli I am OK with having MD5 in there, but it is horribly insecure
and should not be used in production. If we do put it in we should deprecate
it very quickly afterwards and ask people
Github user revans2 commented on the issue:
https://github.com/apache/bookkeeper/pull/128
+1 (non-binding)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so
Github user revans2 commented on the issue:
https://github.com/apache/bookkeeper/pull/126
@eolivelli I totally agree we do need some test cases for this. The reason
that we use the older API is because of the large amount of object allocation
that happens with protocol buffers. I
