Member since
12-01-2016
7
Posts
2
Kudos Received
1
Solution
My Accepted Solutions
Title | Views | Posted |
---|---|---|
995 | 01-11-2017 03:38 PM |
01-05-2018
05:20 PM
A documentation error has been pointed out to us, in the Ambari Installation for Power Systems. (HDP 2.6.3). The HDP-UTILS repo appears to be cited correctly in the "HDP 2.6 Repositories" section within Chapter 1, in the table. However, later in the document, the "Select Version" section of Chapter 3 mentions the repo again but the URL actually points to the tarball rather than the repo itself. (Notice that the Ambari and HDP repos are correct, the HDP-UTILS is wrong). Looking back, I see that this error didn't exist in the prior versions (HDP 2.6.2, 2.6.1, 2.6.0). The latter (ch 3) should be corrected to match the former (from ch 1).
... View more
Labels:
02-20-2017
07:59 PM
Thank you both - I have passed along this information to the client. I've recommended that if they have any trouble getting it working and need any follow-up questions, it would be simpler for them to post their own question in the community (and thus more easily post data about any errors they see, etc).
... View more
02-17-2017
10:30 PM
1 Kudo
A client asks us (thus, this is a "pass-through" question) if they can use MariaDB with Hive on HDP 2.5.3. This case is on Power, on RHEL7, although it doesn't seem like it would be a platform-specific question -- although RHEL7 by default favors mariadb over the earlier, related, mysql. The installation guide identifies mysql-community-java as a prerequisite and it installs this through the HDP-UTILS repo; this subsequently enables the mysql56-community repo. It does refer to documentation for using non-default databases, with a pointer to here: http://docs.hortonworks.com/HDPDocuments/Ambari-2.4.2.0/bk_ambari-reference/content/using_non-default_databases_-_hive.html ... but MariaDB is not included. Can this client use MariaDB with Hive in their HDP 2.5.3 installation? Is there any trick to enabling it?
... View more
Labels:
01-11-2017
03:38 PM
It appears we've discovered the issue. The server has a large number of cores, 160. We have edited /etc/ambari-server/conf/ambari.properties with: agent.threadpool.size.max=128
client.threadpool.size.max=128 (Just increasing to largest ^2 closer to 160); restarting the ambari-server, it appears the clients now register successfully. We should now be able to proceed with the installation.
... View more
01-10-2017
09:27 PM
In a couple attempts to create an HDP 2.5.3 cluster on Power Linux RHEL 7.2, we've seen an error with Ambari-agents connecting to the Ambari server. We have succeeded in creating other clusters with the same versions, including same requisites like openssl versions, openjdk versions, etc. Yet, while many install fine, there is some scenario that prevents success and we have been unable to pin it down. The error is very similar to another reported question, here: https://community.hortonworks.com/questions/24208/help-ambari-agent-registe-fail-with-netutilpy77-er.html However, that one discusses use of Oracle JDK and is on X86 platform. Our case is the same error but on ppc64le. Oracle JDK isn't supported, but OpenJDK 1.8 works (at least on other clusters). Here is some package version output of interest:
[root@st03d ambari-agent]# rpm -qa |grep ambari
ambari-agent-2.4.2.4-3.ppc64le
ambari-server-2.4.2.4-3.ppc64le
[root@st03d ambari-agent]# rpm -qa |grep openjdk |grep 1.8
java-1.8.0-openjdk-headless-1.8.0.65-3.b17.el7.ppc64le
java-1.8.0-openjdk-1.8.0.65-3.b17.el7.ppc64le
java-1.8.0-openjdk-devel-1.8.0.65-3.b17.el7.ppc64le
[root@st03d ambari-agent]# echo $JAVA_HOME
/usr/lib/jvm/java-1.8.0-openjdk
[root@st03d ambari-agent]# rpm -qa |grep openssl
openssl-1.0.1e-51.el7_2.7.ppc64le
openssl-libs-1.0.1e-51.el7_2.7.ppc64le The top of the log looks like this, thru the point of the error:
[root@st03d ambari-agent]# pwd
/var/log/ambari-agent
[root@st03d ambari-agent]# head -15 ambari-agent.log
INFO 2017-01-09 20:59:48,961 main.py:90 - loglevel=logging.INFO
INFO 2017-01-09 20:59:48,961 main.py:90 - loglevel=logging.INFO
INFO 2017-01-09 20:59:48,961 main.py:90 - loglevel=logging.INFO
INFO 2017-01-09 20:59:48,963 DataCleaner.py:39 - Data cleanup thread started
INFO 2017-01-09 20:59:48,965 DataCleaner.py:120 - Data cleanup started
INFO 2017-01-09 20:59:48,966 DataCleaner.py:122 - Data cleanup finished
INFO 2017-01-09 20:59:49,069 PingPortListener.py:50 - Ping port listener started on port: 8670
INFO 2017-01-09 20:59:49,071 main.py:349 - Connecting to Ambari server at https://st03d.pbm.ihost.com:8440 (129.40.28.71)
INFO 2017-01-09 20:59:49,071 NetUtil.py:62 - Connecting to https://st03d.pbm.ihost.com:8440/ca
ERROR 2017-01-09 21:35:22,777 NetUtil.py:88 - EOF occurred in violation of protocol (_ssl.c:765)
ERROR 2017-01-09 21:35:22,777 NetUtil.py:89 - SSLError: Failed to connect. Please check openssl library versions.
Refer to: https://bugzilla.redhat.com/show_bug.cgi?id=1022468 for more details.
WARNING 2017-01-09 21:35:22,777 NetUtil.py:116 - Server at https://st03d.pbm.ihost.com:8440 is not reachable, sleeping for 10 seconds...
INFO 2017-01-09 21:35:32,777 NetUtil.py:62 - Connecting to https://st03d.pbm.ihost.com:8440/ca
WARNING 2017-01-09 21:35:32,778 NetUtil.py:93 - Failed to connect to https://st03d.pbm.ihost.com:8440/ca due to [Errno 111] Connection refused This happens when trying to set up 2-3 nodes, but can also be produced on a single node. Other cluster installs have succeeded, both multi-node and single-node, so the issue is likely something unique to a specific configuration difference that we can't pin down. At the moment we are stumped. What might be wrong here? What else should we look for?
... View more
Labels: