Member since
09-29-2015
32
Posts
41
Kudos Received
6
Solutions
My Accepted Solutions
Title | Views | Posted |
---|---|---|
919 | 01-20-2017 06:15 PM | |
546 | 01-07-2017 12:41 AM | |
855 | 12-20-2016 10:57 PM | |
1472 | 11-05-2015 07:20 PM | |
562 | 09-29-2015 01:56 AM |
02-14-2017
05:42 PM
Perhaps Hive is pointing to the specific NameNode instance and not by the name service ID. You can use "hive metatool" to check (and update the pointer). Please see: https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.4/bk_hadoop-ha/content/ha-update-the-hive-metastore.html
... View more
02-09-2017
06:41 PM
Ah, I see. That's a bummer.
... View more
02-09-2017
06:20 PM
@Clifton Eyre Can you confirm if only the Metric Monitor component is in maintenance mode on that host, or the host itself is in maintenance mode?
... View more
02-09-2017
12:03 AM
Hi Clifton, just to be clear... Are you putting the host in maintenance mode or the specific Metrics Monitor on that host in maintenance mode? If the Metrics Monitor on that host is in maintenance mode but the host is not in maintenance mode, it should technically not affect pushing configs for other components on that host. Also, Metrics Monitor is not part of stack upgrade so the ability to do upgrade should not be affected.
... View more
02-08-2017
11:01 PM
1. Ambari Metrics Monitor is installed by default on all cluster nodes if Ambari Metrics is selected. It cannot be deleted from the UI. It doesn't hurt to leave that on there, but if you really don't want it running on that host, you can go to the Host Details page, shut down Metrics Monitor on that host, and put it in Maintenance Mode, so that alerts are not generated and it won't get started accidentally when doing operations at the Ambari Metrics service level. It's possible to delete it via the API but that's not recommended and you might be on your own because how the UI reacts in that situation isn't really tested (it might get added back automatically, etc). If you are willing to take that risk, you can first stop the Metrics Monitor from the UI on that host and issue the following API call: curl -u admin:admin -H "X-Requested-By: ambari" -X DELETE http://AMBARI_SERVER_HOST:8080/api/v1/clusters/CLUSTERNAME/hosts/HOSTNAME/host_components/METRICS_MONITOR
2. You cannot customize alert thresholds on a per-host basis.
... View more
02-02-2017
02:07 AM
Hi Wataru-san, Gobusatashiteorimasu 🙂 What version of Ambari and HDP are you using? ambarirca database was previously used to store MR job-related info in HDP 1.x. That has been replaced by YARN's App Timeline Service in HDP 2.x. If you are using HDP 2.x, then you don't really need to back that up.
... View more
01-23-2017
09:36 PM
Awesome. If you could "accept" my answer, that would be great. AMS crashing issue might be due to the version of Ambari you are using (2.1.0 is quite old and there have been numerous stability improvements on AMS since then). If it is possible, I highly recommend you upgrade to Ambari 2.4.2.
... View more
01-23-2017
08:00 PM
@Wayne Hoyenga Did the unwanted alerts disappear too?
... View more
01-23-2017
06:54 PM
@Wayne Hoyenga Ok, you just need to issue Stop on that ATS instance on cg-hm09. Please go to Hosts -> cg-hm09 and choose Stop on App Timeline Server from the component list. Then try the delete API call again.
... View more
01-20-2017
11:37 PM
Ok, the ATS instance that you are trying to delete is in one of the states that makes it non-deletable. Can you get me the output of: curl -uadmin:PW -k https://localhost:8443/api/v1/clusters/ROGERGPFS/hosts/cg-hm09.ncsa.illinois.edu/host_components/APP_TIMELINE_SERVER
... View more
01-20-2017
06:48 PM
No worries. If you can provide the stack trace from the ambari-server.log file I should be able to help you further.
... View more
01-20-2017
06:38 PM
🙂 Can you get the corresponding stack trace from the server log? It's at /var/log/ambari-server/ambari-server.log
... View more
01-20-2017
06:15 PM
Thanks for sharing the output. Yes, that's exactly what I meant (REST API call to get the ATS instances registered with Ambari). To delete the bad ATS instance from Ambari, you can issue the following API call: curl -u admin:admin -k -H "X-Requested-By: ambari" -X DELETE https://localhost:8443/api/v1/clusters/ROGERGPFS/hosts/<hostname-with-bad-ATS>/host_components/APP_TIMELINE_SERVER
... View more
01-20-2017
12:20 AM
Hi Wayne. What's the output of /api/v1/clusters/<cluster_name>/host_components?HostRoles/component_name=APP_TIMELINE_SERVER Do you see two ATS instances in the output?
... View more
01-07-2017
12:42 AM
Please refer to http://docs.hortonworks.com/HDPDocuments/Ambari-2.2.2.18/bk_ambari-installation/content/_download_the_ambari_repo.html
... View more
01-07-2017
12:41 AM
1 Kudo
Ambari 2.2.0 did not support HDP 2.4. It was added in Ambari 2.2.1. However, you should use Ambari 2.2.2, as it is more stable.
... View more
12-28-2016
06:00 PM
1 Kudo
You should also be able to specify the condition IsActiveMaster=false to have the API respond with non-active HBaseMasters only. Instead of saying "fields=metrics/hbase/master/IsActiveMaster" in the query, you can turn that into a predicate by saying "metrics/hbase/master/IsActiveMaster=false".
... View more
12-20-2016
11:01 PM
Since Grafana was introduced in Ambari 2.2.2, the sandbox from an older version would not have had the port forwarding rule for Grafana UI that listens on port 3000.
... View more
12-20-2016
10:57 PM
2 Kudos
It seems like the Grafana daemon is up. Is your Virtualbox set up to port forward 3000? If you do "curl http://localhost:3000" from inside the VM's SSH, what do you see?
... View more
04-25-2016
08:57 PM
Hi Saad, what do you mean by that? You do need to install VirtualBox, but you don't really need to do any special configuration for VirtualBox. You just need to follow the instructions in the Quick Start Guide. Let me know if you are stuck at a specific place in the guide. Thanks.
... View more
04-22-2016
06:32 AM
1 Kudo
Elitsa, There are two ways to register hosts with Ambari: automatically via SSH by supplying a private key to Ambari (as you have done) or manually (by running yum install ambari-agent on the host, modifying /etc/ambari-agent/conf/ambari-agent.ini's "hostname" line to the Ambari's hostname, and running "ambari agent start"). When using the SSH method, Ambari does NOT store the SSH private key that you supplied (it is used once to do passwordless SSH into the server to set up the agent and discarded). So there's no way or need to "change the private key" in Ambari.
... View more
04-22-2016
06:16 AM
1 Kudo
Do you know why AppTimelineServer is going down? Is it because of OOM (OutOfMemory)? There was an issue where "yarn-env" configuration deployed by Ambari had an incorrect line for HDP 2.3 so the heap size for AppTimelineServer was always defaulting to 1GB even if you change the AppTimelineServer max heap size through Ambari. Go to Services > YARN > Configs > Advanced > Advanced yarn-env and examine the content of yarn-env template. If the template has this entry: export YARN_HISTORYSERVER_HEAPSIZE={{apptimelineserver_heapsize}} The entry needs to be changed to: export YARN_TIMELINESERVER_HEAPSIZE={{apptimelineserver_heapsize}} Once you make the change, save the config. This problem will be fixed in the upcoming Ambari 2.2.2 via https://issues.apache.org/jira/browse/AMBARI-14715. Not sure if this is the specific issue you are running into.
... View more
04-22-2016
06:00 AM
3 Kudos
If you would like to test installing/managing a multi-node cluster using virtual machines, I would recommend you follow the Ambari Quick Start Guide: https://cwiki.apache.org/confluence/display/AMBARI/Quick+Start+Guide By following the instructions there, you can set up multiple VMs on a single laptop using Vagrant and VirtualBox. I typically use it to create a 3-node cluster. If you have 16GB of RAM, you should be able to comfortably run 3 VMs. I hope this helps.
... View more
01-26-2016
08:36 PM
4 Kudos
Sounds like RHS repos are not available for some reason. I think that's the core issue. You should troubleshoot by running the "yum" command directly from the command line. The template change would disable the repo files distributed by Ambari, but if the RHS repos are not available, then it won't help. If you are manually putting HDP.repo and HDP-UTILS.repo under /etc/yum.repos.d on the cluster hosts, then Ambari will probably overwrite them with the disabled ones... BTW, the repo file that you posted seems to have an error in that "enabled=1" is in the "gpgkey" line (maybe it is just a copy-paste error).
... View more
11-05-2015
08:08 PM
1 Kudo
You can replace sync_type to specific (from existing), and add names attribute with a comma-delimited list of users/groups. Here's an example: curl -uadmin:admin -H 'X-Requested-By: ambari' -X POST -d '[{"Event": {"specs": [{"principal_type": "users", "sync_type": "specific", "names": "bill,jenny,mike"},{"principal_type":"groups","sync_type":"specific", "names": "group1,group2"}]}}]' http://localhost:8080/api/v1/ldap_sync_events
... View more
11-05-2015
07:20 PM
6 Kudos
Try: curl -uadmin:admin -H 'X-Requested-By: ambari' -X POST -d '[{"Event": {"specs": [{"principal_type": "users", "sync_type": "existing"}, {"principal_type": "groups", "sync_type": "existing"}]}}]' http://localhost:8080/api/v1/ldap_sync_events You will get a response like: {
"resources" : [
{
"href" : "http://localhost:8080/api/v1/ldap_sync_events/13",
"Event" : {
"id" : 13
}
}
]
} You can GET on this href to get status of the sync: curl -uadmin:admin http://localhost:8080/api/v1/ldap_sync_events/13
{
"href" : "http://localhost:8080/api/v1/ldap_sync_events/13",
"Event" : {
"id" : 13,
"specs" : [
{
"sync_type" : "existing",
"principal_type" : "users"
},
{
"sync_type" : "existing",
"principal_type" : "groups"
}
],
"status" : "COMPLETE",
"status_detail" : "Completed LDAP sync.",
"summary" : {
"groups" : {
"created" : 0,
"removed" : 0,
"updated" : 0
},
"memberships" : {
"created" : 0,
"removed" : 0
},
"users" : {
"created" : 0,
"removed" : 0,
"updated" : 0
}
},
"sync_time" : {
"end" : 1446751142546,
"start" : 1446751142462
}
}
}
... View more
10-31-2015
08:25 AM
1 Kudo
I would be great if someone can play around and come up with useful and practical Grafana dashboards with the Ambari data source. We can publish these so others can import.
... View more
10-31-2015
08:12 AM
16 Kudos
Grafana is a popular, open-source graph and dashboard builder ( licensed under Apache License, Version 2.0). Grafana is feature rich; not only does it let users create customizable and shareable dashboards, it also offers templated/scripted dashboards, LDAP integration, multiple data sources, and more. Out of the box, Grafana supports data sources such as Graphite, OpenTSDB, ElasticSearch, InfluxDB, CloudWatch, and etc. Luckily, the data source part is pluggable; it was not too difficult to create an Ambari data source for it: https://github.com/u39kun/ambari-grafana It works for all service metrics currently exposed via Ambari: HDFS, YARN, HBase, Storm, Kafka, Flume, Accumulo, and Ambari Metrics. It supports typeahead of component names and their associated metrics, as well as aggregation (“avg”, “sum”, “min”, and “max” of metrics across all instances of the specified component). Host Metrics are not supported yet, but it shouldn't be too difficult to add. Please visit https://github.com/u39kun/ambari-grafana/blob/master/README.md for detailed and easy-to-follow instructions on how to install and use Grafana with Ambari as a data source. Note: Development of the Ambari data source in Grafana was done against Ambari 2.1.2 + HDP 2.3.
... View more
Labels:
10-02-2015
05:52 PM
Did adding smokeuser work, @wgonzalez@hortonworks.com?
... View more
09-29-2015
02:21 AM
1 Kudo
Sounds like https://hortonworks.jira.com/browse/BUG-41647. Does your Hive JDBC URL contain FQDN? If not, YARN config page won't load in Ambari 2.1.0. This was fixed in 2.1.1.
... View more