Member since
02-24-2016
175
Posts
56
Kudos Received
3
Solutions
My Accepted Solutions
| Title | Views | Posted |
|---|---|---|
| 1931 | 06-16-2017 10:40 AM | |
| 16502 | 05-27-2016 04:06 PM | |
| 1632 | 03-17-2016 01:29 PM |
06-02-2016
10:00 AM
@Jonas Straub, @Vipin Rathor : Using this I could export one policy at a time, Is there a way to export all the policies at the same time downloaded using http://<ranger_address>:6080/service/plugins/policies/download/<clustername>_hadoop
... View more
06-02-2016
09:56 AM
@Jonas Straub, @Neeraj Sabharwal, @Sagar Shimpi : Guys, using this I could export the entire policy repository using : <Ranger_Host:IP>/service/public/api/policy and also using http://<ranger_address>:6080/service/plugins/policies/download/<clustername>_hadoop I went through the link shared which talks about exporting policies one by one. Is there a way to export entire repository instead of exporting policies one after another?
... View more
06-01-2016
04:38 PM
@Jonas Straub This is indeed useful. I am able export the policy repositories and policies (by policyId) successfully. However I noticed that exported policy (json) contains ids, now I am trying to create the same policy on the target cluster. However, I do not see the same Id in target db. Mover, I noticed that on source and target ranger db, the service names (i.e. repository names are different), manually I had to change it. But exporting from cluster1 and applying policy on cluster2 worked fine for me. Thank you. Referred documentation here.
... View more
06-01-2016
02:29 PM
Thanks @Sagar Shimpi . Will check this. Meanwhile wanted to know if HWX has plan to have a centralized Ranger (single instance) to define policies for multiple clusters? This could be a great value and help keeping the policies in central location 🙂
... View more
06-01-2016
02:25 PM
@Alex Miller All the users have accounts in both the A/Ds. Except from Cluster-2 's Ranger DB contains users who are no longer in company and in A/D-1, so I am OK scrapping DB of cluster-2. And want to translate policies of A to B and still point to different A/Ds. Of course on the target cluster-2, will stop the user sync module once database is restored from Cluster-A.
... View more
06-01-2016
02:03 PM
1 Kudo
Guys, I have two HDP 2.4.0 clusters. On both the we have setup Ranger and integrated with different ADs. More or less users are same. I read on one of the HCC questions that we can use DB replication to keep both policy DB in sync. What I want to achieve is : From Cluster-A, want to take up Ranger DB, replace environment specific values and restore it on another cluster, restart Ambari and I should be good to go. But more I study schema and spend more time analyzing the situation, it doesn't look practical to me. Has anyone actually replaced DB for Ranger? I was trying to check the Ranger DB and both these environments I found lot of differences in the table contents. Mostly because of the users and groups, as we are pointing to different Active Directories. I've three questions here : (1) If we backup ranger db from cluster-1, drop db of cluster-2, stop user sync in Ambari on cluster-2, restore db of cluster-1 and restart Ambari (keep user sync) off and restart the Ranger admin. Will it work? ( I am happy to have imported users/groups of Cluster-1 - which are imported from AD1 and OK to delete imported users of Cluster-2 Ranger DB, Will this create problem? ) (2) Any better suggestions to make db replication work? (3) What is better way around syncing policies across clusters? Thanks.
... View more
Labels:
- Labels:
-
Apache Ranger
05-27-2016
04:06 PM
1 Kudo
Confirmed with HWX that its a bug.Closing the issue for now.
... View more
05-27-2016
04:04 PM
@Jitendra YadavThanks for confirming. Have raised a case with HWX.
... View more
05-27-2016
09:09 AM
@vshukla You suggesting to change hive.server2.thrift.port in advanced spark-hive-site-override.conf? It is already set to 10015. 1) Did you check on Kerberized cluser? 2) Try setting hive.server2.transport.mode to http .
... View more
05-27-2016
09:03 AM
Yes it is. 10015
... View more