You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ambari.apache.org by Aravindan Vijayan <av...@hortonworks.com> on 2015/11/26 00:12:58 UTC
Review Request 40723: AMBARI-14078 Metrics Collector startup is slow
in Kerberized environment
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/40723/
-----------------------------------------------------------
Review request for Ambari, Dmytro Sen, Sumit Mohanty, and Sid Wagle.
Bugs: AMBARI-14078
https://issues.apache.org/jira/browse/AMBARI-14078
Repository: ambari
Description
-------
PROBLEM
Collector startup takes greater than 6mins even though the ApplicationHistoryServer is up in 2mins in secure cluster.
Bug
The collector startup script uses HBase shell to poll and check if the metrics schema is created. HBase shell takes a long time to respond in a secure environment without kinit.
Fix
Doing a "kinit" before starting up the metrics collector solves the performance issue
-----------------
NOTE : Some metrics collector startup script STDOUT echo statements (4-5) have also been piped to the ambari-metrics-collector.out file to facilitate debugging. Hence, the .out file will be appended to, across collector START requests.
Diffs
-----
ambari-metrics/ambari-metrics-timelineservice/conf/unix/ambari-metrics-collector 25305cf
ambari-server/src/main/resources/common-services/AMBARI_METRICS/0.1.0/package/scripts/ams_service.py 2b475dd
Diff: https://reviews.apache.org/r/40723/diff/
Testing
-------
Manual testing on Kerberized and normal environment.
Python unit tests pass.
Thanks,
Aravindan Vijayan
Re: Review Request 40723: AMBARI-14078 Metrics Collector startup is
slow in Kerberized environment
Posted by Dmytro Sen <ds...@hortonworks.com>.
> On Ноя. 26, 2015, 2:15 д.п., Sid Wagle wrote:
> > ambari-metrics/ambari-metrics-timelineservice/conf/unix/ambari-metrics-collector, line 224
> > <https://reviews.apache.org/r/40723/diff/1/?file=1147125#file1147125line224>
> >
> > Will this command run list * # of tables time? Or will check the output of list run once in a single try against the list of defined tables ?
No it'll run until exit code is 0
if [ $? -eq 0 ]; then
echo "$(date) Ambari Metrics data model initialization completed." | tee -a $OUTFILE
break
fi
It makes a single try against METRICS_AGGREGATE table, which is the last initialized table.
- Dmytro
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/40723/#review108083
-----------------------------------------------------------
On Ноя. 25, 2015, 11:12 п.п., Aravindan Vijayan wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/40723/
> -----------------------------------------------------------
>
> (Updated Ноя. 25, 2015, 11:12 п.п.)
>
>
> Review request for Ambari, Dmytro Sen, Sumit Mohanty, and Sid Wagle.
>
>
> Bugs: AMBARI-14078
> https://issues.apache.org/jira/browse/AMBARI-14078
>
>
> Repository: ambari
>
>
> Description
> -------
>
> PROBLEM
> Collector startup takes greater than 6mins even though the ApplicationHistoryServer is up in 2mins in secure cluster.
>
> Bug
> The collector startup script uses HBase shell to poll and check if the metrics schema is created. HBase shell takes a long time to respond in a secure environment without kinit.
>
> Fix
> Doing a "kinit" before starting up the metrics collector solves the performance issue
>
>
> -----------------
>
> NOTE : Some metrics collector startup script STDOUT echo statements (4-5) have also been piped to the ambari-metrics-collector.out file to facilitate debugging. Hence, the .out file will be appended to, across collector START requests.
>
>
> Diffs
> -----
>
> ambari-metrics/ambari-metrics-timelineservice/conf/unix/ambari-metrics-collector 25305cf
> ambari-server/src/main/resources/common-services/AMBARI_METRICS/0.1.0/package/scripts/ams_service.py 2b475dd
>
> Diff: https://reviews.apache.org/r/40723/diff/
>
>
> Testing
> -------
>
> Manual testing on Kerberized and normal environment.
>
> Python unit tests pass.
>
>
> Thanks,
>
> Aravindan Vijayan
>
>
Re: Review Request 40723: AMBARI-14078 Metrics Collector startup is
slow in Kerberized environment
Posted by Sid Wagle <sw...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/40723/#review108083
-----------------------------------------------------------
ambari-metrics/ambari-metrics-timelineservice/conf/unix/ambari-metrics-collector (line 224)
<https://reviews.apache.org/r/40723/#comment167460>
Will this command run list * # of tables time? Or will check the output of list run once in a single try against the list of defined tables ?
- Sid Wagle
On Nov. 25, 2015, 11:12 p.m., Aravindan Vijayan wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/40723/
> -----------------------------------------------------------
>
> (Updated Nov. 25, 2015, 11:12 p.m.)
>
>
> Review request for Ambari, Dmytro Sen, Sumit Mohanty, and Sid Wagle.
>
>
> Bugs: AMBARI-14078
> https://issues.apache.org/jira/browse/AMBARI-14078
>
>
> Repository: ambari
>
>
> Description
> -------
>
> PROBLEM
> Collector startup takes greater than 6mins even though the ApplicationHistoryServer is up in 2mins in secure cluster.
>
> Bug
> The collector startup script uses HBase shell to poll and check if the metrics schema is created. HBase shell takes a long time to respond in a secure environment without kinit.
>
> Fix
> Doing a "kinit" before starting up the metrics collector solves the performance issue
>
>
> -----------------
>
> NOTE : Some metrics collector startup script STDOUT echo statements (4-5) have also been piped to the ambari-metrics-collector.out file to facilitate debugging. Hence, the .out file will be appended to, across collector START requests.
>
>
> Diffs
> -----
>
> ambari-metrics/ambari-metrics-timelineservice/conf/unix/ambari-metrics-collector 25305cf
> ambari-server/src/main/resources/common-services/AMBARI_METRICS/0.1.0/package/scripts/ams_service.py 2b475dd
>
> Diff: https://reviews.apache.org/r/40723/diff/
>
>
> Testing
> -------
>
> Manual testing on Kerberized and normal environment.
>
> Python unit tests pass.
>
>
> Thanks,
>
> Aravindan Vijayan
>
>