Member since
09-24-2015
816
Posts
488
Kudos Received
189
Solutions
My Accepted Solutions
Title | Views | Posted |
---|---|---|
2687 | 12-25-2018 10:42 PM | |
12286 | 10-09-2018 03:52 AM | |
4218 | 02-23-2018 11:46 PM | |
1915 | 09-02-2017 01:49 AM | |
2237 | 06-21-2017 12:06 AM |
12-15-2016
12:57 PM
Thanks for the update!
... View more
12-10-2016
06:23 AM
Hi @Ron Ballard, the list of encodings depends on what's installed on your machines. You can list them using for example "iconv -l". CP1252 is installed on Hortonworks Sandbox and has another 2 aliases: $ iconv -l | grep 1252
CP1252 MS-ANSI WINDOWS-1252 Regarding your second question, the answer is Yes, and that's exactly what I'm doing above. "city" is using the default UTF-8 encoding and city_sjis is in SJIS: create table city as select * from city_sjis;
... View more
11-18-2016
11:10 AM
1 Kudo
Can you try to create the keytab for the test user, and always do kinit from the keytab ("kinit -kt <keytab-file-name>" command). Ambari service checks are supposed to use the ambari-qa user and the wizard is supposed to create its keytab as well.
... View more
10-22-2016
10:39 AM
It seems that something is blocking pre-emption. For example, if Spark jobs ask for large executors, "total_preemption_per_round" or another such property may block their release. See here for details how pre-emption works. Try to reduce spark queue max capacity to 75% or so, or talk to the Dev team and ask them to cooperate by using smaller executors, like 1G. Also set the memory allocated for Yarn containers to be a multiple of min container size.
... View more
10-10-2016
02:01 AM
You are missing kernel-headers. Check /etc/yum.conf file for a line saying "exclude=kernel*" and comment it out. Then, as a test run "yum install kernel-headers" and "yum install hadoop_2_5_0_0_1245" on one node. If it works, comment out that line on all cluster nodes and retry the install by Ambari. More details about kernel headers here.
... View more
10-10-2016
01:49 AM
1 Kudo
Adding multiple Knox gateways to respective cluster nodes by Ambari is supported, and has been supported for a while, see details here. You need to set the LB by yourself. Also, you need to sync {GATEWAY_HOME}/conf/security/ folders over all Knox instances, using one of them as a source; this is not done by Ambari.
... View more
10-10-2016
01:21 AM
1 Kudo
It refers to active Name Node. For distcp you are providing 2 paths, source and target files/directories, both on HDFS of respective clusters. With some more settings you can refer to NN name service, meaning that you don't need to care which NN is active.
... View more
10-09-2016
02:14 AM
1 Kudo
How about yarn.resourcemanager.monitor.capacity.preemption.max_wait_before_kill ? The default is 15 sec (15000 ms), see here and here for details.
... View more
10-09-2016
01:39 AM
Shiro Provider is used to authenticate users, when they enter user-name/password against an LDAP store. See here for details. "main.ldapRealm.contextFactory.url" must be $protocol :// $ldaphost : $port, where $protocol is ldap or ldaps pointing to a node which provides LDAP service like OpenLDAP, AD, FreeIPA, or the test LDAP server bundled with Knox, running on port 33389. KDC doesn't provide LDAP service.
... View more
10-06-2016
08:22 AM
You may have some network issues, if I understand the log correctly it is saying that the assigned Application master for your spark job cannot access the driver running on your external node. Can cluster nodes access your external node on ephemeral port numbers like 33656?
... View more