> On July 31, 2017, 4:25 p.m., Alejandro Fernandez wrote: > > ambari-server/src/main/java/org/apache/ambari/server/controller/internal/ClusterStackVersionResourceProvider.java > > Lines 1123 (patched) > > <https://reviews.apache.org/r/61239/diff/1/?file=1785542#file1785542line1123> > > > > Add a safeguard here in case indexOf returns -1. Only if > 0 should the > > substring method be called to get the os. > > This also assumes majorVersion is a single digit, instead look for the > > numeric characters from the end. > > Di Li wrote: > Hello Alejandro, > > The osFamily var was checked earlier in the code if "-ppc" exists. So by > the time it reaches the indexOf, the osFamily can either be null (already > safeguarded) or a string that contains "-ppc", so the indexof will have > positive value. > > For single digit issue, the current logic can deal with multiple digit > """ > String osType = "redhat100"; > String os = osType.substring(0, "redhat-ppc".indexOf("-ppc")); > String majorVersion = osType.substring(os.length()); > String result = String.format("OS: %s, majorVersion: %s", os, > majorVersion); > System.out.println(result); > """ > OUTPUT: OS: redhat, arch: ppc, majorVersion: 100 > > Alejandro Fernandez wrote: > This is good programming practice. If the caller of the function changes, > then this will raise an exception.
Hello Alejandro, The code is updated. please review again, thx. - Di ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/61239/#review181811 ----------------------------------------------------------- On July 31, 2017, 6:20 p.m., Di Li wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/61239/ > ----------------------------------------------------------- > > (Updated July 31, 2017, 6:20 p.m.) > > > Review request for Ambari, Alejandro Fernandez, Jonathan Hurley, Sid Wagle, > and Tim Thorpe. > > > Bugs: AMBARI-21611 > https://issues.apache.org/jira/browse/AMBARI-21611 > > > Repository: ambari > > > Description > ------- > > Error message: Repositories for os type redhat-ppc7 are not defined. > This is because: > AmbariAgent reports os type (to be stored in the Hosts table) as redhat-ppc7. > "-ppc" is a hardcoded post fix attached to the os type in os_check.py > """ > os_check.py (AMBARI-21054) > if _is_powerpc(): > operatingSystem += '-ppc' > """ > When Ambari server backend generates the > {os, repo} > map, it uses os information from repo_versoin as the key . But later on uses > os info from the Hosts table to query the map, thus the mismatch (redhat-ppc7 > vs redhat7). > The repo_version table repository column however only uses redhat6/7 and does > not distinguish ppc vs x86 ( nor it should ) > > > Diffs > ----- > > > ambari-server/src/main/java/org/apache/ambari/server/controller/internal/ClusterStackVersionResourceProvider.java > 633fe8c > > ambari-server/src/test/java/org/apache/ambari/server/controller/internal/ClusterStackVersionResourceProviderTest.java > 7018245 > > > Diff: https://reviews.apache.org/r/61239/diff/2/ > > > Testing > ------- > > one IOP ppc one IOP x86, upgrade Ambari, patch with server jar, run install > packages after registering HDP 2.6.2 stack. verify the error did not appear > any more for ppc, both ppc and x86 cluster were able to finish the install > package step successfully > > > Thanks, > > Di Li > >
