incubator-tashi-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael Stroucken <...@cmu.edu>
Subject Re: IPMC release vote on Apache Tashi 201202-incubating
Date Mon, 05 Mar 2012 01:57:38 GMT
Michael Stroucken wrote:
> Our two active mentors have voted +1 on our release proposal, but 
> three votes were needed. I think we should probably try to recruit a 
> replacement mentor. While I don't think projects under incubation are 
> required to have three mentors, the project is not of general enough 
> interest to get other people with binding votes to check it out.
Paul Fremantle has motivated ant elder to examine the release candidate, 
and he gave his +1. That means our first release is approved!

Now to move all the bits to the right place...
>
> At this point, I'd feel better rolling a new release incorporating the 
> bugfixes and improvements of the past few weeks.
We the developers have now been granted sufficient access to the 
standard Apache JIRA issue tracking system. We've added new bugs that we 
observed, closed ones that were fixed in the past, and fixed several 
issues that will not be available in this release.

These are:-
TASHI-2: fix indentations (repeat): unknown how much this affected Tashi 
deployers, or how many new bugs this fix will uncover. Whitespace 
sensitivity FTW!
TASHI-8: shutdownMany command added to client
TASHI-10: friendlier output to tashi-client mistakes
TASHI-11: prevent two interfaces being connected to the same network
TASHI-15: let Tashi daemons exit on SIGINT

(new development): let connections become re-established when a handle 
can no longer be communicated over. There was functionality in the CM 
and NM before the major rewrite in December, but this improvement is now 
available to any RPC using part of Tashi.
>
> There currently is no change in wire protocol; the new shutdownMany 
> function is implemented via the client. Therefore, users of the 
> current stable branch or release candidate can do a staggered deployment.
Yes, no changes in wire protocol, so a rolling upgrade should be possible.
>
> From what I see on other votes on iao, it may be possible to propose 
> another release candidate to iao directly, without having to call for 
> a vote here again. Is my understanding of this correct?
I've been informed that we will have to do the full release vote program 
over again. I'll start that in the next few days.

Greetings,
Michael.


Mime
View raw message