> On June 30, 2015, 12:14 a.m., Adam B wrote: > > docs/persistent-volume.md, line 56 > > <https://reviews.apache.org/r/35981/diff/1/?file=994064#file994064line56> > > > > Can a volume/reservation created by one framework principal only be > > destroyed/unreserved by a framework scheduler using the same principal? If > > the principal is different, then the operation would fail, even if the role > > or even frameworkId is identical?
We only look at the principal of an framework or an operator. The role does not matter because an operator does not have a role. The frameworkId is also irrelavent because framework B can unreserve/destroy a reservation/volume created by framework B as long as ACL allows it. > On June 30, 2015, 12:14 a.m., Adam B wrote: > > docs/persistent-volume.md, lines 67-69 > > <https://reviews.apache.org/r/35981/diff/1/?file=994064#file994064line67> > > > > You might need a blank line preceding this, and/or some indentation for > > it to show up as a numbered list. Doesn't render properly as is. > > > > Unique per role per slave? So I could use the same volumeId for related > > volume instances (e.g. my 'logging' or 'data' volumes), up to a max of one > > per slave? But if I happen to launch another executor/task on a slave where > > one of those volumes is already created/mounted, then I have to rename it > > ('data2')? Yes, this is in the design doc. Persistent volumes can be shared within a role, so it has to be unique per role. > On June 30, 2015, 12:14 a.m., Adam B wrote: > > docs/persistent-volume.md, lines 120-122 > > <https://reviews.apache.org/r/35981/diff/1/?file=994064#file994064line120> > > > > Any chance it could return the host_path too? Or do I have to go > > digging through slave logs to figure out where my volume was created? Why the framework wants to know that? This is similar to sandbox. Do we have a way to let the framework know about the location of the sandbox? - Jie ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/35981/#review89827 ----------------------------------------------------------- On July 7, 2015, 12:34 p.m., Michael Park wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/35981/ > ----------------------------------------------------------- > > (Updated July 7, 2015, 12:34 p.m.) > > > Review request for mesos, Adam B, Jie Yu, and Timothy Chen. > > > Bugs: MESOS-2405 > https://issues.apache.org/jira/browse/MESOS-2405 > > > Repository: mesos > > > Description > ------- > > The Github rendered version is available [here]( > https://github.com/mesosphere/mesos/blob/user-docs/docs/persistent-volume.md) > > > Diffs > ----- > > docs/persistent-volume.md PRE-CREATION > > Diff: https://reviews.apache.org/r/35981/diff/ > > > Testing > ------- > > Documentation. > > > Thanks, > > Michael Park > >
