That's what I was thinking... retiring indicates not just that it's not currently being developed, but that it can't be developed if needed. It makes sense for projects to drop it, because they will have no way to patch it if there's an issue. And, with something like HTrace, which one embeds deep in one's code, projects need to plan ahead for that contingency, rather than use it until there's an issue.
On Sat, Mar 10, 2018 at 11:42 AM Billie Rinaldi <billie.rina...@gmail.com> wrote: > I feel like the htrace core library doesn't really need active development, > but I worry that retiring htrace will encourage other projects to drop it > as a dependency. If it had a home as a subproject, perhaps people would > continue to use it even if releases were infrequent. IMO that would be > preferable to retirement. I'd be on board if a TLP wanted to adopt htrace > as a subproject, but I also expect I will continue to only have time for > votes and not development. > > On Thu, Mar 8, 2018 at 2:00 PM, Christopher <ctubb...@apache.org> wrote: > > > On Thu, Mar 8, 2018 at 4:52 PM Todd Lipcon <t...@cloudera.com> wrote: > > > > > On Thu, Mar 8, 2018 at 1:38 PM, Mike Drob <md...@apache.org> wrote: > > > > > > > +1 to retire. > > > > > > > > This is painful because I've spent a lot of time on this personally, > > and > > > I > > > > do think there is value, but not if nobody picks it up. > > > > > > > > > > > > Christopher, > > > > > > > > HTrace did not originate from Accumulo's cloudtrace effort. It spun > out > > > of > > > > HDFS's efforts. > > > > > > > > They were two separate things, where eventually cloudtrace was > replaced > > > by > > > > HTrace. Look at https://issues.apache.org/jira/browse/ACCUMULO-898 > > > > > > > > > Actually going back further than that, it did actually come from > > > cloudtrace. It was an intern that I and a few others worked with at > > > Cloudera many summers back who did the porting out of cloudtrace to > make > > it > > > a more generic thing in a separate community. > > > > > > -Todd > > > > > > > In retrospect, it is a shame more Accumulo folks didn't help out making > it > > a more generic thing once it moved to Incubator. It's probably too late > > now, but I wouldn't mind helping out with maintaining HTrace and pushing > it > > towards graduation, if there's still others willing to stay involved (ASF > > requires 3 minimum). I've already been helping downstream, packaging > HTrace > > for Fedora, and would like to see it continue. Does anybody know what > will > > happen to HTrace if/when this vote passes? I'm not sure, but I don't > think > > there's a lot of interest in rebooting cloudtrace in Accumulo, but maybe > it > > will revert to its former GitHub presence? Is there a current plan? > > >
