Hide Forgot
Description of problem: With only the basic set of pcp packages installed, we ran "sanity" group of upstream tests. Some of them reported "not run" because of missing requirements (e.g. TC 369: [not run] Perl PCP::LogImport module is not installed), and such behavior makes sense to me. pcp-pmda-news was *not* installed, and TC 416 produced hard fail: [79%] 416 - output mismatch (see 416.out.bad) :: [ BEGIN ] :: Running 'diff -urN /var/lib/pcp/testsuite/416.out /var/lib/pcp/testsuite/416.out.bad' --- /var/lib/pcp/testsuite/416.out 2016-01-15 11:15:32.000000000 -0500 +++ /var/lib/pcp/testsuite/416.out.bad 2016-01-26 13:03:07.881151012 -0500 @@ -1,66 +1,2 @@ QA output created by 416 -dbpmda> open pipe PERL PCP_PMDAS_DIR/news/pmdanews.pl -Start perl PMDA: PERL PCP_PMDAS_DIR/news/pmdanews.pl -dbpmda> traverse news -Metric: news - news.articles.count - news.articles.last - news.articles.total - news.readers.nnrpd - news.readers.rn - news.readers.trn - news.readers.vn - news.readers.xrn -dbpmda> children news.articles -Metric: news.articles - leaf count - leaf last - leaf total -dbpmda> children news.articles.count -Metric: news.articles.count -dbpmda> pmid news.articles.total -Metric: news.articles.total - 28.0.201 -dbpmda> fetch news.articles.total -PMID(s): 28.0.201 -pmResult dump from ADDR timestamp: 0.000000 TIMESTAMP numpmid: 1 - 28.0.201 (news.articles.total): numval: 1 valfmt: 0 vlist[]: - value 800 1.1210388e-42 0x320 -dbpmda> desc news.articles.count -PMID: 28.0.301 - Data Type: 32-bit unsigned int InDom: 28.0 0x7000000 - Semantics: instant Units: count -dbpmda> instance 28.0 -pmInDom: 28.0 -[ 0] inst: 1 name: "comp.sys.sgi" -[ 1] inst: 2 name: "comp.sys.sgi.graphics" -[ 2] inst: 3 name: "comp.sys.sgi.hardware" -[ 3] inst: 4 name: "sgi.bad-attitude" -[ 4] inst: 5 name: "sgi.engr.all" -dbpmda> fetch news.articles.count -PMID(s): 28.0.301 -pmResult dump from ADDR timestamp: 0.000000 TIMESTAMP numpmid: 1 - 28.0.301 (news.articles.count): numval: 5 valfmt: 0 vlist[]: - inst [1 or ???] value 0 0 0x0 - inst [2 or ???] value 0 0 0x0 - inst [3 or ???] value 0 0 0x0 - inst [4 or ???] value 700 9.8090893e-43 0x2bc - inst [5 or ???] value 100 1.4012985e-43 0x64 -dbpmda> profile 28.0 none -dbpmda> profile 28.0 add 2 -dbpmda> profile 28.0 add 4 -dbpmda> getdesc on -dbpmda> fetch news.articles.count -PMID(s): 28.0.301 -pmResult dump from ADDR timestamp: 0.000000 TIMESTAMP numpmid: 1 - 28.0.301 (news.articles.count): numval: 2 valfmt: 0 vlist[]: - inst [2 or ???] value 0 - inst [4 or ???] value 700 -dbpmda> close -dbpmda> quit - -=== news.log === -Log for pmdanews on ... - - -Log finished ... +Where is /var/lib/pcp/pmdas/news? Would it make sense to "fix" TC 416 that it won't run when news PMDA is not available? I suppose it does eventually find out ("Where is /var/lib/pcp/pmdas/news?"), would it be possible to realize such situation earlier, and gracefully quit as "not run"? Version-Release number of selected component (if applicable): Only this set of pcp packages was installed. pcp-3.10.9-5.el6.ppc64 pcp-conf-3.10.9-5.el6.ppc64 pcp-libs-3.10.9-5.el6.ppc64 pcp-libs-devel-3.10.9-5.el6.ppc64 pcp-testsuite-3.10.9-5.el6.ppc64 python-pcp-3.10.9-5.el6.ppc64 How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Hey Miloš, Thanks for catching this, and your suggestion is precisely correct, we need a 'notrun' test checking for the actual pmdanews.pl itself. I've added this commit to my upstream branch (should be picked up in master eventually). Should be easy to backport if required. ec24477151b7e0737b8fea2d24d9e585d91c3831 qa/416 add notrun install test for the pmda itself Previously we only checked that perl bindings were installed. With the pmda subrpms, we should also check that the pmda itself is installed
Verified for build pcp-3.10.9-6.el6.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2016-0825.html