You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ambari.apache.org by Andrew Onischuk <ao...@hortonworks.com> on 2015/08/25 14:20:02 UTC

Review Request 37758: Ambari startup scripts exhibit long delays (5+ seconds each) on every "modifying user" operation. This adds a few minutes to the time it takes to start or restart any component, causing excessive downtime for every config change.

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37758/
-----------------------------------------------------------

Review request for Ambari and Vitalyi Brodetskyi.


Bugs: AMBARI-12871
    https://issues.apache.org/jira/browse/AMBARI-12871


Repository: ambari


Description
-------

PROBLEM: Customer use nslcd to fetch the user from AD server, it returns a lot
of users and groups, this will slow down the ambari operation (start/stop
service). As discussed with Florian Barca, The root cause is a limitation in
the native API: grp.getgrall() has no ability to prefilter the results, which
means it fetches all users from the LDAP directory exposed by nslcd. If it
happen to be a large org, with a large number of users, so the fetch takes a
long time. We need to optimize this and make the code scale with the number of
users.


Diffs
-----

  ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py 699fe5f 
  ambari-server/src/test/python/stacks/2.0.6/hooks/before-ANY/test_before_any.py 129907f 

Diff: https://reviews.apache.org/r/37758/diff/


Testing
-------

mvn clean test


Thanks,

Andrew Onischuk