> On Dec. 7, 2016, 8:47 p.m., Kevin Klues wrote: > > src/slave/containerizer/mesos/io/switchboard.cpp, line 1102 > > <https://reviews.apache.org/r/54496/diff/1/?file=1579072#file1579072line1102> > > > > We never launch the io switchboard process with vlog turned on, so we > > may never see this log. But it doesn't hurt to have during debugging.
As per offline discussion with Vinod and Kevin, let me think about how we want to go about this i.e., - If we need logging here at all (since anything we log goes to to the STDERR of the container) - If we want to invoke `await(...)` and then log. Marking the issue as fixed for now. - Anand ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/54496/#review158391 ----------------------------------------------------------- On Dec. 7, 2016, 8:15 p.m., Anand Mazumdar wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/54496/ > ----------------------------------------------------------- > > (Updated Dec. 7, 2016, 8:15 p.m.) > > > Review request for mesos, Jie Yu, Kevin Klues, and Vinod Kone. > > > Bugs: MESOS-6746 > https://issues.apache.org/jira/browse/MESOS-6746 > > > Repository: mesos > > > Description > ------- > > The switchboard never used to wait for the read end of the pipe > to finish processing the data before exiting. This was problematic > as the client connected to the output entrypoint might not receive > the data. Note that we are relying on `socket.close()` to > eventually ensure that the data is flushed to the client when the > socket is closed. > > > Diffs > ----- > > src/slave/containerizer/mesos/io/switchboard.cpp > 4452d5ad26881487825e58a3a052e6d07bac2321 > > Diff: https://reviews.apache.org/r/54496/diff/ > > > Testing > ------- > > make check (A test is added later in the chain that won't pass otherwise) > > > Thanks, > > Anand Mazumdar > >
