----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/21498/#review43345 -----------------------------------------------------------
Is it possible that someone is relying on the existing behaviour? I realise that you've made a new method to mimic the existing behaviour, but that will still break people's scripts unless they update them to call the new method. My bias tends to be to preserve the API until the next official revision, and work around the problem on the client side. - Stephen Turner On May 15, 2014, 7:54 p.m., Sam Schmit wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/21498/ > ----------------------------------------------------------- > > (Updated May 15, 2014, 7:54 p.m.) > > > Review request for cloudstack. > > > Bugs: CLOUDSTACK-6009 > https://issues.apache.org/jira/browse/CLOUDSTACK-6009 > > > Repository: cloudstack-git > > > Description > ------- > > CLOUDSTACK-6009: listHosts API call is returning memoryAvailable and > memoryTotal in Bytes, and memoryUsed in Kilobytes. This fix changes the > memoryUsed value to return in Bytes as well, and includes a new method to > return memoryUsed in Kilobytes if needed. > > > Diffs > ----- > > api/src/com/cloud/host/HostStats.java 4eb7b1a > core/src/com/cloud/agent/api/GetHostStatsAnswer.java 6a52e76 > core/src/com/cloud/agent/api/HostStatsEntry.java c9d25a0 > > Diff: https://reviews.apache.org/r/21498/diff/ > > > Testing > ------- > > Pre-change: > 1) Called "listHosts" API call with no arguments. > 2) Validated that memoryAvailable and memoryTotal were in Bytes, while > memoryUsed was in Kilobytes (factor of 1000 times smaller than it should be) > > Post-change: > 1) Called "listHosts" API call with no arguments. > 2) Validated that memoryAvailable, memoryTotal, and memoryUsed are all in > Bytes, and memoryAvailable + memoryUsed = memoryTotal > > > Thanks, > > Sam Schmit > >
