----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/7642/#review12578 -----------------------------------------------------------
src/master/master.hpp <https://reviews.apache.org/r/7642/#comment26848> Or more simply: Information about a connected or completed framework. src/messages/messages.proto <https://reviews.apache.org/r/7642/#comment26849> See comment below about executor run. src/slave/slave.hpp <https://reviews.apache.org/r/7642/#comment26850> The problem is that a run id can be reused. So, over the course of some framework, it can have an executor with run 0 and uuid == 'aa' and later run 0 and uuid == 'bb' because the initial run directory got cleaned up. This is making me think that the correct strategy is to ditch run numbers all together and look up executors with these uuids. In fact, given this review, I'm pretty sure if an executor with id say 'default' were to die, and another one came up in it's place we would not be able to get to the sandbox of 'default/runs/0' but only 'default/runs/1' (that is, we won't actually be able to see old runs!). This is going to be the common case, so we definitely need to handle it. - Benjamin Hindman On Oct. 18, 2012, 1:49 a.m., Ben Mahler wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/7642/ > ----------------------------------------------------------- > > (Updated Oct. 18, 2012, 1:49 a.m.) > > > Review request for mesos, Benjamin Hindman and Vinod Kone. > > > Description > ------- > > -The slave now stores completed frameworks, executors, and tasks in memory. > -They are also exported via the state JSON endpoint. > -The webui has been updated accordingly. > > This is branched off of vinod's change here: > https://reviews.apache.org/r/7638/ > > Note that the JSON response sizes grow significantly due to this change, I'll > be following up with a change to pull in zlib and compress our HTTP Responses. > > > Diffs > ----- > > src/master/constants.hpp 4372f4d9d40b21701feae83ab56dc18afa735899 > src/master/http.cpp 9254859a0d4fddf775fd38d0482b4622dc767ab3 > src/master/master.hpp 146af017bbb6da9bd44acb53a4f1ee0ffbedd64f > src/master/master.cpp 82e4dc704e5c67ec178bd058934896328308d868 > src/messages/messages.proto 4e0538fe929f9091e5cdd4a1bb017d836df52a3e > src/slave/constants.hpp 4e28d30ebe3490f555a6dd43da9c6c1374ea4883 > src/slave/http.cpp d413c3c1bd9b6772d5ebf23fa8ae1bd5ae3870bb > src/slave/slave.hpp 343c353027dbd2a7c1be8cee99a1d59367169177 > src/slave/slave.cpp 5af7464aae17c00a0e707421982d7cb055aabc6c > src/webui/master/static/browse.html > e7e09a28ca9893ba049af100a257631d427ebfa1 > src/webui/master/static/controllers.js > 02a045d4a2ac01d468adc0885097f9191d7f1f08 > src/webui/master/static/framework.html > ecbf879b217ab51ea6f2531ebadd8dc09fd8ca54 > src/webui/master/static/frameworks.html > 4da01583eeb5bcd6ee5a36d4cbd58042bafcab15 > src/webui/master/static/home.html 84e9317b1b3bc857d6d6532adb25fd21a7691732 > src/webui/master/static/slave.html b96dfe95e08e79f683c622bf0bec96b9342c3be1 > src/webui/master/static/slave_executor.html > 33fa5f4ba3bacb5bb73849c1b1fa76a417351aa1 > src/webui/master/static/slave_framework.html > b43bbdcd79fab62d28e8d53737e3c4a083e0812d > > Diff: https://reviews.apache.org/r/7642/diff/ > > > Testing > ------- > > OSX + CentOS: make check > local run with long-lived-framework > > > Thanks, > > Ben Mahler > >
