Thanks Heng. Do you know if 1.2.2 or earlier versions of 1.2 failed in similar way or is this new phenomenon? Thanks, St.Ack
On Mon, Sep 5, 2016 at 1:26 AM, Heng Chen <heng.chen.1...@gmail.com> wrote: > - Unpacked source and binary tarballs: layout looks good > - Started up a 3-node cluster (Hadoop 2.7.2, Oracle JDK 8u20, 2 master, 2 > rs) from binary tarballs. > - Verified that the web UI works and shell works > - build from source and run test case (JDK 8u20), passed. > - Run LTT with 1M rows (100 writers, 30 readers (100%), 10 updaters > (20%)) all keys verified, no warns, no errors, no failed, latencies > lgtm > - Run ITBLL with 2M rows (slowDeterministic), passed. > > Run ITBLL with 2M rows (serverKilling) has some issues. I run two times, > all failed ( hbase --config ~/hadoop/hbase/f04_conf/ > org.apache.hadoop.hbase.test.IntegrationTestBigLinkedList loop 1 1 > 2000000 /tmp/it_16224_v0 1 -m serverKilling) > > The first time, after kill components in cluster serval times (1 active > master, 1 backup master, 1 active RS, 1 dead RS), some regions fall in > Failed_open state on active RS, so balance could not run, and ITBLL hang. > Then i start the dead RS manually, and stop the active RS, ITBLL could > go on. But after some time, ITBLL hang again due to backup master could > not be startup, so i do manually again, and ITBLL go on , and at least > verified failed. > > > The second time, there is no abnormal issues during ITBLL, but verified > failed. > > Upload the logs > > > > 2016-09-03 0:33 GMT+08:00 Misty Stanley-Jones <mi...@apache.org>: > >> +1 based on OSX 10.11.6 >> >> Steps taken: >> >> Binary tar.gz: >> - Download the tarball >> - Test the MD5sum, it matched >> - Extract the tarball >> - Start HBase in standalone mode >> - Start the CLI >> - Access the master and regionServer web UIs >> - Stop HBase >> >> Source tar.gz: >> - Download the tarball >> - Test the MD5sum, it matched >> - Extract the tarball >> - Build using Maven 3.39 and JDK 1.8.0_102 on OSX and 'mvn clean install >> --fail-at-end' and let the full test suite run >> -- A few failed tests but each passed when I ran it separately. Not >> surprising since I was running all this on a Macbook. >> - Start HBase in standalone mode >> - Start the CLI >> - Access the master and regionServer web UIs >> - Stop HBase >> >> >> On Tue, Aug 30, 2016, at 11:14 AM, Stack wrote: >> > The first release candidate for HBase 1.2.3 (hbase-1.2.3RC0) is >> > available for download at: >> > >> > https://dist.apache.org/repos/dist/dev/hbase/hbase-1.2.3RC0/ >> > >> > Maven artifacts are also available in a staging repository at: >> > >> > https://repository.apache.org/content/repositories/orgapache >> hbase-1149/ >> > >> > Artifacts are signed with my key (30CD0996) published up in our KEYS >> > file at https://www-us.apache.org/dist/hbase/KEYS. >> > >> > The RC is tagged 1.2.3RC0 (I'll sign the tag the next time through...) >> > >> > The detailed source and binary compatibility report vs 1.2.2 has been >> > published for your review, at: >> > >> > http://people.apache.org/~stack/1.2.2_1.2.3RC0_compat_report.html >> > >> > HBase 1.2.3 is the third patch release in the HBase 1.2 line, continuing >> > on >> > the theme of bringing a stable, reliable database to the Hadoop and >> NoSQL >> > communities. This release includes over 48 bug fixes since 1.2.2. In >> > particular it addresses an API incompatibility uncovered by our Apache >> > Phoenix bothers and sisters in our Table Interface: HBASE-16420. >> > >> > The full list of fixes included in this release is available at: >> > >> > >> > https://issues.apache.org/jira/secure/ReleaseNote.jspa?proje >> ctId=12310753&version=12336053 >> > >> > and in the CHANGES.txt file included in the distribution. >> > >> > Please try out this candidate and vote +/-1 by 23:59 Pacific time on >> > Monday, 2016-09-05 as to whether we should release these artifacts as >> > HBase >> > 1.2.3. >> > >> > Thanks, >> > St.Ack >> > >