* fix recordio split and task passes

* update for pre commit

* update

* update, still need to sync client wait for pass end.

* able to sync passes for task dispatching

* update to comment

* update

* fix yapf check

* why local pre-commit fails? version is the same

* fix race condition

* update

* fix race condition

* this still have duplicate problem in unit test

* update

* update

* update by comment

* update

Unchecked errors could be handled by: cd go; gometalinter --vendor --disable-all --enable errcheck $(glide nv)

* using etcd as fault tolerant training

* update

* workable version, ft not tested

* small fix

* update

* remove TODO

* init etcd cclient

* add etcd

* add etcd.go

* fix compile problem

* move code to etcd.go

* add etcd_lister.go for pserver client

* add etcd_client_test.go

* merge etcd_client_test and client_test

* refine client_test.go

* refine code

* format code

* add TODO and use interface instead of struct

* fix typo of initDesiredPservers

* optimize dir structure of go/pserver/client

* add a flag to config index for pserver

* follow comment

* fix path

* optimize code

* remove err in pserver NewEtcd

* restore comment about /ps_desired

* add trainer library

* modifty file name

* move trainer to master client

* update

* update

* modify monitor master to receive a chan

* update

* use etcd client from etcd_client.go

* update

* update

* remove etcd client without lock

* update

* update the comment

* update commonts

The pserver need to have server index for saving model on the
distributed file system. The server index comes from etcd if etcd is
used, or user can manually specify them. So we need
pserver.NewService() to take index as an argument. Since index could
come from etcd, it would be strange if pserver takes an index as
argument, at the same time get the index from etcd. so we will need to
decouple pserver with etcd.