About jeden

jeden · ‎05-03-2016

I got it. So the documented behavior is accurate if the client is not directly running on a DataNode, but does not account for WebHDFS or a client directly connected to the node. Thanks!

jeden · ‎05-03-2016

When writing files larger than one block, how are blocks distributed across the datanodes? Documentation seems to indicate large files are split across datanodes (whenever possible), but I'm not sure this is always the case.

jeden · ‎02-03-2016

The API is not an option at this point. If there's no way to do this in Ambari natively, that's fine - I just need to know that. But workarounds of any kind do not meet the need.

jeden · ‎02-03-2016

🙂 For purposes of this question, yes.

jeden · ‎02-03-2016

Do we have documentation available around Ambari Flume management? Based on what I know so far, this is what would make sense to me in the real world, assuming an admin does not want to be involved in setting up / maintaining / managing Flume agents: 1) Set up a separate Ambari Server instance for Flume management. (Treat Flume Configs tab like a View, in essence... if that's doable?) Thus, Flume "admins" (developers) don't have access to HDP cluster management via Ambari. 2) Add a single HDP edge node (logically speaking, at least) to the Ambari instance (Is it possible for two Ambari instances to manage different services on the same node?), as well as all other servers on which Flume agents are to be installed via Ambari add hosts. 3) Create config groups based on server identity - ex: all web servers might have identical log tracking, the HDP edge node(s) would be its (their) own config group, etc. I don't have a test environment robust enough to try this out, so any help is appreciated. Thanks!

jeden · ‎01-25-2016

@Neeraj Sabharwal Interesting. So the command you used *did* allow you to set node labels on the fly. If I'm reading correctly, the main difference was the ResourceRequest.setNodeLabelExpression nodeLabel being replaced with -node_label_expression node1 Is that right? I went back and checked the documentation, and saw that it now matches what you performed. So the core of my issue in particular was a bug in the docs. Thanks for following up!

jeden · ‎01-15-2016

Not really. I have a coworker going back to validate what I saw now. Assuming he sees the same things I did, we'll be checking to see if bugs have been filed, etc. Neeraj, anything to add on your part? @Mark Herring @Neeraj Sabharwal

jeden · ‎12-15-2015

Ahhh, perfect. Thanks!

jeden · ‎12-15-2015

Here's part 2 of the answer, with the screenshot showing my current proxyuser settings for falcon and oozie: I realize it's possible I've missed something...

jeden · ‎12-15-2015

Balu - here are the results of the cURL from the Falcon server -- it works fine: As far as the log file goes, the error I'm getting back indicates failure because Falcon is not set up as a proxy user. Here's the pertinent snippet, I think: However, as I will show in my next post, I have set up both falcon and oozie proxyuser settings as * so I'm not sure why that would be an issue here.

jeden · ‎12-15-2015

I am attempting to create a Falcon cluster entity using the Web UI. **My Environment** I have a functioning Oozie Server (have successfully run workflows from the command line, and I can access the Oozie Web UI without issue). I have a three-node cluster running in Docker. Both Falcon and Oozie are installed on node2. **The Error** Note: The way name resolution is set up, FQDNs are not required in my environment, or rather, "node2" is the FQDN equivalent. **Troubleshooting So Far** I have tried replacing node2 with the IP address. I get the same error: I have confirmed that I can access this URL via the Oozie Web UI. Just to be pedantic, I also performed a port scan to confirm that port 11000 was truly open on the server. What could be causing this error? Are there additional troubleshooting steps I can take?

jeden · ‎11-30-2015

Steve - that's a good point. The question, then, is why make it available as a cluster-level service at all? What is the benefit of providing Slider as an installable option, or providing a Slider view in HDP?

jeden · ‎11-21-2015

Ah, gotcha. So installing the Slider View is functionally equivalent to installing Slider on the cluster, assuming you are launching applications from the View. But to make slider available via CLI, you must install it at the cluster level. Makes sense!

jeden · ‎11-20-2015

Successfully running application, Slider View: List of installed services, no Slider:

jeden · ‎11-20-2015

I am running HDP 2.3 and was capturing screenshots for a lesson under development for Slider. I created a Slider view, created the yarn home directory in HDFS, copied an HBase slider package to the appropriate location, and restarted Ambari server. It was then I realized the I had forgotten to install the Slider service on my cluster. Thinking it would be a great error message to capture, I went back to the Slider View, created the app per the wizard, and deployed - thinking it would error out and tell me that the service was not installed. To my surprise, the application ran without a problem, and the Slider View allowed me to perform the Flex and other actions with no issues. Because this is for a training class, I need to understand why this worked. Is there something special about the HBase slider package we make available that allows it to run on a cluster where the service is not installed, or does the HDP or Ambari support slider applications without the need to explicitly install the service? If that's true, what is it that installing Slider on HDP actually accomplishes? I will post screenshots in a comment below. Thanks in advance.

jeden · ‎11-19-2015

Thank you!

jeden · ‎11-19-2015

Here's the exact command and error returned. Also, here's the documentation link: http://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.2/bk_yarn_resource_mgt/content/ref-635c245f-d56c-49c0-97b7-a789012ea71d.1.html

jeden · ‎11-19-2015

According to our docs, I should be able to download a Slider package for HBase at the following link: http://s3.amazonaws.com/dev.hortonworks.com/HDP/centos6/2.x/updates/2.3.2.0/slider-app-packages/hbase/slider-hbase-app-package-1.1.1.2.3.2.0-2950.zip However, I am getting a "forbidden" error when I attempt to access/download anything. Is there a reason for this? Is this available at an alternative location?

jeden · ‎11-11-2015

That was it! Thanks Jonas!

jeden · ‎11-11-2015

And remember, I am writing a publicly available training course, so please, no answers that require hacking that you would not suggest the average administrator to attempt. 🙂

jeden · ‎11-11-2015

UPDATE: I finally ran across this issue, which at least explains why the process is not working: https://issues.apache.org/jira/browse/AMBARI-12811 The question now, then, becomes "How do I update the repo_version table prior to a blueprint-based installation?"

jeden · ‎11-11-2015

I have successfully uploaded a blueprint to an Ambari server. I am trying now to configure default repositories for HDP installation in our classroom lab environment, as per instructions here: https://cwiki.apache.org/confluence/display/AMBARI/Blueprints#Blueprints-Step4:SetupStackRepositories%28Optional%29 I have created the a file, named HDPRepo.test, per the instructions. The command I am using is based on what worked for the blueprint upload, but pointed at the repo URI per the instructions: curl -u admin:admin -i -H "X-Requested-By: jedentest1" -X POST -d @HDPRepo.test http://node1:8080/api/v1/stacks/HDP/versions/2.3/operating_systems/redhat6/repositories/HDP-2.3/ I can't find this demo'd anywhere on the web, so I'm flying a bit blind. The error being returned is: { "status" : 500, "message" : "Cannot create repositories."} My suspicion is that the documentation is missing some minor piece (possibly a cURL add-on, or perhaps the API URI isn't quite right) that makes this update of the base_url value work. Has anyone actually done this that can point out what I need to modify in my command? Thanks!

jeden · ‎11-04-2015

Sure! Let me know if you want a peek at my setup.

jeden · ‎11-04-2015

Background: I have configured two node labels - HiCPU (exclusive=false) and GPU (exclusive=true). I have attached HiCPU to a queue named Engineering with 100% capacity. Label GPU is attached to a queue named Marketing with 100% capacity. No default label has been configured for either queue at the beginning of the test. Issue 1: When I run the following commands as the hdfs user, the command will only run on an unlabeled node, and if no unlabeled nodes are available, the job simply hangs: yarn jar /usr/hdp/current/hadoop-yarn-client/hadoop-yarn-applications-distributedshell.jar -shell_command "sleep 25" -jar /usr/hdp/current/hadoop-yarn-client/hadoop-yarn-applications-distributedshell.jar -num_containers 3 -queue Engineering ResourceRequest.setNodeLabelExpression HiCPU or yarn jar /usr/hdp/current/hadoop-yarn-client/hadoop-yarn-applications-distributedshell.jar -shell_command "sleep 25" -jar /usr/hdp/current/hadoop-yarn-client/hadoop-yarn-applications-distributedshell.jar -num_containers 3 -queue Marketing ResourceRequest.setNodeLabelExpression GPU However, if I set the node label as default, the commands *DO* execute on the appropriate machine, even without the ResourceRequest.setNodeLabelExpression attribute (as would be expected). Bottom line - I can only get node labels to work for a YARN job if they are set as the default, which means non-labeled nodes are not available to that queue any longer for YARN jobs. Issue 2: Our documentation here: http://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.2/bk_yarn_resource_mgt/content/using_node_labels.html ...states the following: "...if you submit a MapReduce job to a queue that has a default node label expression, the default node label will be applied to the MapReduce job." To test this, I executed the following command using a user who was default-queue-mapped to the Engineering queue: yarn jar /usr/hdp/2.3.0.0-2557/hadoop-mapreduce/hadoop-mapreduce-examples.jar pi 5 10 While the job did get assigned to the correct queue, in no instance could I get the MapReduce job to run on a labeled node. It would run on an unlabeled node if available, and if no unlabeled nodes were available, it would just hang. Thus, unless I am missing something, it appears that the *only* way to get a functional use of node labels is to set them as a default for a queue, in which case YARN jobs assigned to that queue will ONLY run on the labeled nodes (including unlabeled YARN jobs). Furthermore, under no circumstance will a MapReduce job run on a labeled node, regardless of default node label settings. If someone wants to take a look at my settings on my cluster and troubleshoot, let me know. Thanks! @Neeraj

jeden · ‎10-29-2015

Also - is there a quicker / easier / cleaner way to test the node labels that you used other than the one listed in the docs?

jeden · ‎10-29-2015

Again, with caveats. For example, the sample command given to test the settings says to run it as yarn, however it failed for me as some permissions now apparently required the hdfs user. My bottom line, which would be nice to validate, is this: 1) The command line setup of users and folders is unnecessary, and Ambari sets them up slightly differently than our docs 2) Permissions are also different for folders (yarn hadoop vs. yarn yarn...) 3) The sample hadoop jar command to test (which I ran as yarn jar) requires switching to hdfs rather than yarn. Does that sound about right?

jeden · ‎10-29-2015

Does anyone have (or can you point me to) a link that explains the steps of using YARN node labels in HDP 2.3 being managed by Ambari 2.1? Our existing documentation is for manual configuration only. I've clicked the enabled button, and have already figured out that Ambari creates slightly different directory structures, etc. than are listed. I can keep going with trial and error, but if someone has already documented the process, caveats, etc. that would be fantastic. Thanks!

Online	Offline
Last Visited	‎10-07-2016 03:49 PM

Date Registered	‎09-10-2015 08:30 PM
Last Visited	‎10-07-2016 03:49 PM
Total Messages Posted	27
Total Kudos Received	9

Re: Ambari Flume Management Best Practices - Docum...

Re: HDFS File Placement when File Size Exceeds Blo...

HDFS File Placement when File Size Exceeds Block S...

Re: Ambari Flume Management Best Practices - Docum...

Re: Ambari Flume Management Best Practices - Docum...

Ambari Flume Management Best Practices - Documenta...

Re: YARN Node Labels Not Working ?? Two Issues

Re: YARN Node Labels Not Working ?? Two Issues

Re: Troubleshoot Falcon Web UI Cluster Creation - ...

Re: Troubleshoot Falcon Web UI Cluster Creation - ...

Re: Troubleshoot Falcon Web UI Cluster Creation - ...

Troubleshoot Falcon Web UI Cluster Creation - Inva...

Re: Slider service install not required for slider...

Re: Slider service install not required for slider...

Re: Slider service install not required for slider...

Slider service install not required for slider app...

Re: Access to HBase for Slider Packages?

Re: Access to HBase for Slider Packages?

Access to HBase for Slider Packages?

Re: Use cURL to make an API call to change base UR...

Re: Use cURL to make an API call to change base UR...

Re: Use cURL to make an API call to change base UR...

Use cURL to make an API call to change base URL re...

Re: YARN Node Labels Not Working ?? Two Issues

YARN Node Labels Not Working ?? Two Issues

Re: Node Label Configuration with HDP 2.3.x and Am...

Re: Node Label Configuration with HDP 2.3.x and Am...

Node Label Configuration with HDP 2.3.x and Ambari...