RD, You can take a look at JobHistoryUtils in org.apache.hadoop.mapreduce.v2.jobhistory. This might be helpful for programmatic api to get the jobhistory information.
mahadev On Sat, Dec 3, 2011 at 12:37 PM, RD <rdsr...@gmail.com> wrote: > For Parsing job history logs in H23, One way I see is to figure out the > history file path > (JobHistoryUtils.getConfiguredHistoryServerDoneDirPrefix(conf)/YYYY/MM/DD/<job > serial number>/<job_id>-*.jhist) . > Then parse it with JobHistoryParser to get a JobInfo object out of it. > > There's also a JobHistory server now, but I don't see code which would > return me jobhistory file location or some parseable content (json?) > for constructing the Jobinfo obj. > > I was just curious that whether the first approach is OK, or are there more > better ways to achieve this? > >
