> On 2012-03-27 22:43:05, Matei Zaharia wrote: > > This looks good, except for a few minor things: > > 1) If you just run these out of the box, the slave will silently fail to > > start because its "--master" command line flag is not passed. We should > > tell the user to configure this in mesos.conf, or try to set it > > automatically if there is only one slave. > > 2) The previous scripts attempted to automatically find the amount of cores > > and memory available. Should we add that back? > > 3) It's not clear that the master and slaves are logging to anywhere (if > > you just run /sbin/mesos-master for example, it says "Logging to" and does > > not give a path). If they aren't, we should update the docs to say that you > > can turn on logging by setting log_dir. Ditto for the work directory (where > > will that be after a make install? It seems like the default is ./work). > > Benjamin Hindman wrote: > 1) I've updated the README with a reminder about at least the "--master" > flag. > 2) No. I'd rather not do anything implicitly. As long as we are forcing > users to update configuration files, they might as well add this information > in themselves. How difficult this will be for them to do (or the lack of > being able to do this programmatically) is a different story, but I don't > want to tackle that now, before the release. If you have suggestions for this > please create a JIRA where we can brainstorm together! > 3) I've updated the output to say "Logging to <stderr>". I've also moved > the work_dir to default to /tmp/mesos (this is okay for multiple slaves > because each actual work directory for the slaves will be under > /tmp/mesos/slaves/ID/...). > > Matei Zaharia wrote: > Regarding 2, why not add back the old code (the one that set extra > options in > https://github.com/apache/mesos/blob/trunk/src/deploy/mesos-daemon)? Why > break existing functionality for the release? > > Regarding 3, just mention that you need to set log_dir to see output when > using the deploy scripts, because by default stderr gets redirected to > /dev/null. > > Benjamin Hindman wrote: > 2) We are breaking all sorts of functionality for these deploy scripts, > for example, no more local mesos-env.sh being used to set options remotely. > If the plan was to go with simplicity, and we're already breaking > functionality, why add these back? > 3) Okay, I'll add this to the README.
Well, it's a matter of judgement, but I would say that anything that saves the user's time is worth it. My own goal with the scripts was not "simplicity" for the sake of simplicity. It was to make Mesos as easy to use as possible. Anyone setting it up on a cluster will probably want to use all the CPUs and memory present on each machine, so why not do that by default? Again, this is probably not obvious at Twitter, but Mesos is far from easy to set up and manage for people who haven't done it before. We should really strive to have a setup procedure that can be described in 4-5 shell commands. Each additional step turns people off. The Hadoop tutorial was a great example of doing that, but we need a similar thing for setting up your first cluster. It doesn't need to be the most feature-complete way to manage a cluster; advanced users might switch to monit or chef or something anyway; but it needs to save the user time and give them a reasonable running system with minimal effort. - Matei ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/4491/#review6455 ----------------------------------------------------------- On 2012-03-27 23:12:32, Benjamin Hindman wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/4491/ > ----------------------------------------------------------- > > (Updated 2012-03-27 23:12:32) > > > Review request for mesos, Andy Konwinski and Matei Zaharia. > > > Summary > ------- > > Updated deploy scripts to work with latest version of Mesos. Eliminated > 'deploy-to-slaves' functionality, instead we assume the binaries and > libraries will get installed via other means. Updated the semantics of where > configuration options are gathered from since a single configuration file was > insufficient given conflicting configuration options between binaries (e.g., > mesos-master and mesos-slave). Added a section to the README describing these > scripts. > > > Diffs > ----- > > README 8166b04 > configure.ac 7a0d89f > src/Makefile.am 448410e > src/conf/deploy-env.sh.template a4565e2 > src/configurator/configurator.hpp ecd1365 > src/configurator/configurator.cpp 1f8134c > src/deploy/deploy-to-slaves 0e10486 > src/deploy/mesos-daemon b3144c0 > src/deploy/mesos-daemon.sh.in PRE-CREATION > src/deploy/mesos-deploy-env.sh.template PRE-CREATION > src/deploy/mesos-env.sh 14ad67a > src/deploy/mesos-start-cluster.sh.in PRE-CREATION > src/deploy/mesos-start-masters.sh.in PRE-CREATION > src/deploy/mesos-start-slaves.sh.in PRE-CREATION > src/deploy/mesos-stop-cluster.sh.in PRE-CREATION > src/deploy/mesos-stop-masters.sh.in PRE-CREATION > src/deploy/mesos-stop-slaves.sh.in PRE-CREATION > src/deploy/start-masters a450ee0 > src/deploy/start-mesos 9e5b922 > src/deploy/start-slaves b3b931f > src/deploy/stop-masters f2aa9e2 > src/deploy/stop-mesos 679a09e > src/deploy/stop-slaves 141acd4 > src/master/main.cpp e88315f > > Diff: https://reviews.apache.org/r/4491/diff > > > Testing > ------- > > N/A > > > Thanks, > > Benjamin > >
