Skip to content

durability for regression #648

kumagi opened this Issue Feb 5, 2014 · 4 comments

2 participants

Jubatus member
kumagi commented Feb 5, 2014

now, classifier get a durability for machine failure.
the same mechanism is needed for regression too.

@kumagi kumagi was assigned Feb 5, 2014
@kmaehashi kmaehashi added this to the Far Future milestone Feb 10, 2014
@kmaehashi kmaehashi modified the milestone: Near Future, Far Future Mar 3, 2014
@kumagi kumagi modified the milestone: Near Future Mar 3, 2014
@kmaehashi kmaehashi modified the milestone: 0.7.1, Near Future Mar 2, 2015
Jubatus member
kumagi commented Mar 23, 2015

does regression use simple_storage ?
if so, durability is inherited from classifier?

@kmaehashi kmaehashi modified the milestone: 0.7.1, 0.7.2 Apr 3, 2015
Jubatus member

@kumagi As commented in #648 (comment), this issue may be already fixed? If so, please close this issue.

@kmaehashi kmaehashi assigned kmaehashi and unassigned kumagi Apr 27, 2015
@kmaehashi kmaehashi modified the milestone: 0.8.0, 0.7.2 Apr 27, 2015
Jubatus member

I confirmed that this feature is already implemented for regression too.
When a new server tries to join to the existing cluster, the model is automatically copied from existing server.

2015-06-09 13:03:19,465 17317 INFO  [linear_mixer.cpp:395] start to get model from other server
2015-06-09 13:03:19,467 17202 INFO  [linear_mixer.cpp:571] sending learning-model. size = 191
@kmaehashi kmaehashi closed this Jun 9, 2015
@kmaehashi kmaehashi removed this from the 0.8.0 milestone Jun 9, 2015
Jubatus member

Cleared milestone.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.