Re: Review Request 48622: AMBARI-17136: If Solr is down or not ready, then LogFeeder to should retry

2016-06-13 Thread Oliver Szabo 

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/48622/#review137274
---


Ship it!




Ship It!

- Oliver Szabo


On June 13, 2016, 9:28 a.m., Don Bosco Durai wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/48622/
> ---
> 
> (Updated June 13, 2016, 9:28 a.m.)
> 
> 
> Review request for Ambari, Dharmesh Makwana, Miklos Gergely, Oliver Szabo, 
> and Sumit Mohanty.
> 
> 
> Bugs: AMBARI-17136
> https://issues.apache.org/jira/browse/AMBARI-17136
> 
> 
> Repository: ambari
> 
> 
> Description
> ---
> 
> Uploaded new patch based on feedback from previous review board requests
> 
> 
> Diffs
> -
> 
>   
> ambari-logsearch/ambari-logsearch-logfeeder/src/main/java/org/apache/ambari/logfeeder/output/OutputSolr.java
>  6fb0b0e 
> 
> Diff: https://reviews.apache.org/r/48622/diff/
> 
> 
> Testing
> ---
> 
> Tested on 3 node hadoop cluster
> 
> 
> Thanks,
> 
> Don Bosco Durai
> 
>

Re: Review Request 48622: AMBARI-17136: If Solr is down or not ready, then LogFeeder to should retry

2016-06-13 Thread Miklos Gergely 

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/48622/#review137273
---


Ship it!




Ship It!

- Miklos Gergely


On June 13, 2016, 9:28 a.m., Don Bosco Durai wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/48622/
> ---
> 
> (Updated June 13, 2016, 9:28 a.m.)
> 
> 
> Review request for Ambari, Dharmesh Makwana, Miklos Gergely, Oliver Szabo, 
> and Sumit Mohanty.
> 
> 
> Bugs: AMBARI-17136
> https://issues.apache.org/jira/browse/AMBARI-17136
> 
> 
> Repository: ambari
> 
> 
> Description
> ---
> 
> Uploaded new patch based on feedback from previous review board requests
> 
> 
> Diffs
> -
> 
>   
> ambari-logsearch/ambari-logsearch-logfeeder/src/main/java/org/apache/ambari/logfeeder/output/OutputSolr.java
>  6fb0b0e 
> 
> Diff: https://reviews.apache.org/r/48622/diff/
> 
> 
> Testing
> ---
> 
> Tested on 3 node hadoop cluster
> 
> 
> Thanks,
> 
> Don Bosco Durai
> 
>

Re: Review Request 48622: AMBARI-17136: If Solr is down or not ready, then LogFeeder to should retry

2016-06-13 Thread Don Bosco Durai 


> On June 13, 2016, 9:56 a.m., Miklos Gergely wrote:
> > ambari-logsearch/ambari-logsearch-logfeeder/src/main/java/org/apache/ambari/logfeeder/output/OutputSolr.java,
> >  line 354
> > 
> >
> > addRouterField should be called outside of the loop. The purpose of 
> > this function is to distribute the uploaded documents evenly among the 
> > shards, so it is not important to ensure that the ROUTER_FIELD is 
> > calculated based on the actual sending time, there is no need to re-set it 
> > in every iteration. The addRouterField function may not throw an exception 
> > so it is not a problem if it is not in any try-catch block.
> > 
> > Using setField instead of addField is a good decision, it describes 
> > better what the program's intention is.

In a multi-node SolrCloud cluster, each shard might be in different nodes. I 
have seen in some cases, if particular shard has any long running issue, then 
it is better to move on to next shard so that we don't block forever.


- Don Bosco


---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/48622/#review137260
---


On June 13, 2016, 9:28 a.m., Don Bosco Durai wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/48622/
> ---
> 
> (Updated June 13, 2016, 9:28 a.m.)
> 
> 
> Review request for Ambari, Dharmesh Makwana, Miklos Gergely, Oliver Szabo, 
> and Sumit Mohanty.
> 
> 
> Bugs: AMBARI-17136
> https://issues.apache.org/jira/browse/AMBARI-17136
> 
> 
> Repository: ambari
> 
> 
> Description
> ---
> 
> Uploaded new patch based on feedback from previous review board requests
> 
> 
> Diffs
> -
> 
>   
> ambari-logsearch/ambari-logsearch-logfeeder/src/main/java/org/apache/ambari/logfeeder/output/OutputSolr.java
>  6fb0b0e 
> 
> Diff: https://reviews.apache.org/r/48622/diff/
> 
> 
> Testing
> ---
> 
> Tested on 3 node hadoop cluster
> 
> 
> Thanks,
> 
> Don Bosco Durai
> 
>

Re: Review Request 48622: AMBARI-17136: If Solr is down or not ready, then LogFeeder to should retry

2016-06-13 Thread Miklos Gergely 

---
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/48622/#review137260
---




ambari-logsearch/ambari-logsearch-logfeeder/src/main/java/org/apache/ambari/logfeeder/output/OutputSolr.java
 (line 341)


addRouterField should be called outside of the loop. The purpose of this 
function is to distribute the uploaded documents evenly among the shards, so it 
is not important to ensure that the ROUTER_FIELD is calculated based on the 
actual sending time, there is no need to re-set it in every iteration. The 
addRouterField function may not throw an exception so it is not a problem if it 
is not in any try-catch block.

Using setField instead of addField is a good decision, it describes better 
what the program's intention is.


- Miklos Gergely


On June 13, 2016, 9:28 a.m., Don Bosco Durai wrote:
> 
> ---
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/48622/
> ---
> 
> (Updated June 13, 2016, 9:28 a.m.)
> 
> 
> Review request for Ambari, Dharmesh Makwana, Miklos Gergely, Oliver Szabo, 
> and Sumit Mohanty.
> 
> 
> Bugs: AMBARI-17136
> https://issues.apache.org/jira/browse/AMBARI-17136
> 
> 
> Repository: ambari
> 
> 
> Description
> ---
> 
> Uploaded new patch based on feedback from previous review board requests
> 
> 
> Diffs
> -
> 
>   
> ambari-logsearch/ambari-logsearch-logfeeder/src/main/java/org/apache/ambari/logfeeder/output/OutputSolr.java
>  6fb0b0e 
> 
> Diff: https://reviews.apache.org/r/48622/diff/
> 
> 
> Testing
> ---
> 
> Tested on 3 node hadoop cluster
> 
> 
> Thanks,
> 
> Don Bosco Durai
> 
>