----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/39417/#review104119 -----------------------------------------------------------
docs/configuration.md (lines 1545 - 1551) <https://reviews.apache.org/r/39417/#comment162363> Can we not infer where to put the classifier from any exisiting classifiers and only require the user to specify its parent if they want particular behavior? src/slave/containerizer/isolators/network/port_mapping.cpp (line 398) <https://reviews.apache.org/r/39417/#comment162359> s/Failed tokenize/Failed to tokenize/ src/slave/containerizer/isolators/network/port_mapping.cpp (lines 401 - 405) <https://reviews.apache.org/r/39417/#comment162360> Why not add a 0x prefix if it's not present so you can use numify? if (!strings::startsWith(tokens[0], "0x")) { ... } src/slave/containerizer/isolators/network/port_mapping.cpp (lines 407 - 409) <https://reviews.apache.org/r/39417/#comment162361> This will get refactored as above, but it's more helpful to specifically state which token couldn't be parsed. src/slave/containerizer/isolators/network/port_mapping.cpp (line 1415) <https://reviews.apache.org/r/39417/#comment162364> kill this newline so the error checking is paired with the Try<> src/slave/containerizer/isolators/network/port_mapping.cpp (line 1419) <https://reviews.apache.org/r/39417/#comment162365> add a newline in here src/slave/containerizer/isolators/network/port_mapping.cpp (lines 1423 - 1424) <https://reviews.apache.org/r/39417/#comment162366> Can we verify this somehow? What's the failure mode? src/slave/flags.cpp (lines 506 - 513) <https://reviews.apache.org/r/39417/#comment162362> Why is this description different from the docs? - Ian Downes On Oct. 17, 2015, 5:29 p.m., Cong Wang wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/39417/ > ----------------------------------------------------------- > > (Updated Oct. 17, 2015, 5:29 p.m.) > > > Review request for mesos, Ian Downes and Jie Yu. > > > Repository: mesos > > > Description > ------- > > When --egress_unique_flow_per_container is enabled, we need to install a flow > classifier (fq_codel) qdisc on egress side. This flag specifies where to > install it in the hierarchy. By default, we install it at root. But, for > example, if you have already installed HTB qdisc at root, you may want this > to be installed in other place than root, specify an HTB class ID as its > parent here. > > > Diffs > ----- > > docs/configuration.md 9443d5fc2de0f46f9e117b08e29b09ff3a4579c6 > src/slave/containerizer/isolators/network/port_mapping.cpp > e6bb75e6f5ba48a0c4cf6dd8f353e5f5843d0eef > src/slave/flags.hpp 3e93b52a5874f52361d5a9c685499a7032014a73 > src/slave/flags.cpp 1bf394ea62fde29caa6705cd5d156eae452adbf2 > > Diff: https://reviews.apache.org/r/39417/diff/ > > > Testing > ------- > > Manual tests, with and without a pre-installed HTB qdisc and classes. > > > Thanks, > > Cong Wang > >
