[jira] [Commented] (OAK-5368) Not configurable/Unnecessary short Lucene Observation Queue length

2017-03-08 Thread Stefan Eissing (JIRA)

[ 
https://issues.apache.org/jira/browse/OAK-5368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15901046#comment-15901046
 ] 

Stefan Eissing commented on OAK-5368:
-

You are correct. I must have scanned for creations of {{BackgroundObserver}} 
and found the {{LuceneIndexProviderService}} that, line 411, 
{{registerObserver}} with a fixed length of 5.

For my tests with hitting queue length max, I increased this and then observed 
the measurements attached to this thread. With all changes, I did no longer see 
the warnings described and claimed to have fixed something which, in this case, 
was never there.

I do not see any other problems with event compaction in this Lucene observer. 
So, from my pov: sorry about the noise. This ticket can be resolved as 
worksforme.

> Not configurable/Unnecessary short Lucene Observation Queue length
> --
>
> Key: OAK-5368
> URL: https://issues.apache.org/jira/browse/OAK-5368
> Project: Jackrabbit Oak
>  Issue Type: Improvement
>  Components: lucene
>Affects Versions: 1.4.10
>Reporter: Stefan Eissing
>Assignee: Chetan Mehrotra
>Priority: Minor
> Fix For: 1.8
>
> Attachments: LuceneIndexConfigObserver-under-load.png
>
>
> The maximum queue length in the {{LuceneIndexConfigObserver}} is hard coded 
> to 5. This is unreasonable short for production systems experiencing heavy 
> load.
> Tests with a patched version resulted in observed queue length of >100 
> entries. This floods warnings into the error log, produces additional load.
> The fix would be to increase the queue max, make it configurable or rely on 
> the system default (which can be configured).



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)


[jira] [Commented] (OAK-5368) Not configurable/Unnecessary short Lucene Observation Queue length

2017-03-08 Thread Chetan Mehrotra (JIRA)

[ 
https://issues.apache.org/jira/browse/OAK-5368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15901002#comment-15901002
 ] 

Chetan Mehrotra commented on OAK-5368:
--

ChangeProcessor is used for JCR Listerners so it would not be involved with 
Observer registered by Lucene

> Not configurable/Unnecessary short Lucene Observation Queue length
> --
>
> Key: OAK-5368
> URL: https://issues.apache.org/jira/browse/OAK-5368
> Project: Jackrabbit Oak
>  Issue Type: Improvement
>  Components: lucene
>Affects Versions: 1.4.10
>Reporter: Stefan Eissing
>Assignee: Chetan Mehrotra
>Priority: Minor
> Fix For: 1.8
>
> Attachments: LuceneIndexConfigObserver-under-load.png
>
>
> The maximum queue length in the {{LuceneIndexConfigObserver}} is hard coded 
> to 5. This is unreasonable short for production systems experiencing heavy 
> load.
> Tests with a patched version resulted in observed queue length of >100 
> entries. This floods warnings into the error log, produces additional load.
> The fix would be to increase the queue max, make it configurable or rely on 
> the system default (which can be configured).



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)


[jira] [Commented] (OAK-5368) Not configurable/Unnecessary short Lucene Observation Queue length

2017-03-08 Thread Stefan Eissing (JIRA)

[ 
https://issues.apache.org/jira/browse/OAK-5368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15901000#comment-15901000
 ] 

Stefan Eissing commented on OAK-5368:
-

[~chetanm] unfortunately I do not have the logs any longer. However it was the 
warning that is given by the {{ChangeProcessor}} (trunk line 332) if I remember 
correctly. Does that help?

> Not configurable/Unnecessary short Lucene Observation Queue length
> --
>
> Key: OAK-5368
> URL: https://issues.apache.org/jira/browse/OAK-5368
> Project: Jackrabbit Oak
>  Issue Type: Improvement
>  Components: lucene
>Affects Versions: 1.4.10
>Reporter: Stefan Eissing
>Assignee: Chetan Mehrotra
>Priority: Minor
> Fix For: 1.8
>
> Attachments: LuceneIndexConfigObserver-under-load.png
>
>
> The maximum queue length in the {{LuceneIndexConfigObserver}} is hard coded 
> to 5. This is unreasonable short for production systems experiencing heavy 
> load.
> Tests with a patched version resulted in observed queue length of >100 
> entries. This floods warnings into the error log, produces additional load.
> The fix would be to increase the queue max, make it configurable or rely on 
> the system default (which can be configured).



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)


[jira] [Commented] (OAK-5368) Not configurable/Unnecessary short Lucene Observation Queue length

2017-03-07 Thread Chetan Mehrotra (JIRA)

[ 
https://issues.apache.org/jira/browse/OAK-5368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15900768#comment-15900768
 ] 

Chetan Mehrotra commented on OAK-5368:
--

[~stefan.eissing] From code flow I do not see any warning logging enabled when 
queue becomes full. Can you attach the logs which you saw related to this 
observer?

> Not configurable/Unnecessary short Lucene Observation Queue length
> --
>
> Key: OAK-5368
> URL: https://issues.apache.org/jira/browse/OAK-5368
> Project: Jackrabbit Oak
>  Issue Type: Improvement
>  Components: lucene
>Affects Versions: 1.4.10
>Reporter: Stefan Eissing
>Assignee: Chetan Mehrotra
>Priority: Minor
> Fix For: 1.8
>
> Attachments: LuceneIndexConfigObserver-under-load.png
>
>
> The maximum queue length in the {{LuceneIndexConfigObserver}} is hard coded 
> to 5. This is unreasonable short for production systems experiencing heavy 
> load.
> Tests with a patched version resulted in observed queue length of >100 
> entries. This floods warnings into the error log, produces additional load.
> The fix would be to increase the queue max, make it configurable or rely on 
> the system default (which can be configured).



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)


[jira] [Commented] (OAK-5368) Not configurable/Unnecessary short Lucene Observation Queue length

2016-12-23 Thread Stefan Eissing (JIRA)

[ 
https://issues.apache.org/jira/browse/OAK-5368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15772404#comment-15772404
 ] 

Stefan Eissing commented on OAK-5368:
-

[~chetanm] I agree that it will work for the compacted changes and it always 
recovers quite fast in my tests. If it has performance impact I cannot say, 
since I investigate other, larger performance blocks. I mainly want to clear 
the error logs from things that are not relevant warnings and give a notice 
that a length of 5 is not realistic for a busy server.

I think, if it weren't for reusing the BackgroundObserver, a queue length of 1 
would be fine where the producer always replaces any existing change with a 
new, info=null instance. Which goes into queue compaction strategies,

For this ticket, if we could eliminate the unnecessary warning, I'd be good.

> Not configurable/Unnecessary short Lucene Observation Queue length
> --
>
> Key: OAK-5368
> URL: https://issues.apache.org/jira/browse/OAK-5368
> Project: Jackrabbit Oak
>  Issue Type: Improvement
>  Components: lucene
>Affects Versions: 1.4.10
>Reporter: Stefan Eissing
>Assignee: Chetan Mehrotra
>Priority: Minor
> Fix For: 1.8
>
> Attachments: LuceneIndexConfigObserver-under-load.png
>
>
> The maximum queue length in the {{LuceneIndexConfigObserver}} is hard coded 
> to 5. This is unreasonable short for production systems experiencing heavy 
> load.
> Tests with a patched version resulted in observed queue length of >100 
> entries. This floods warnings into the error log, produces additional load.
> The fix would be to increase the queue max, make it configurable or rely on 
> the system default (which can be configured).



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (OAK-5368) Not configurable/Unnecessary short Lucene Observation Queue length

2016-12-22 Thread Chetan Mehrotra (JIRA)

[ 
https://issues.apache.org/jira/browse/OAK-5368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15771723#comment-15771723
 ] 

Chetan Mehrotra commented on OAK-5368:
--

[~stefan.eissing] The low queue size is by design as this Observer is not 
interested in distinguishing between local and external change. So on queue 
full if it gets compacted then also its fine. The observer does precise diff 
based on loaded index in IndexTracker. 

Are you seeing any performance impact in diff due to low queue size?

> Not configurable/Unnecessary short Lucene Observation Queue length
> --
>
> Key: OAK-5368
> URL: https://issues.apache.org/jira/browse/OAK-5368
> Project: Jackrabbit Oak
>  Issue Type: Improvement
>  Components: lucene
>Affects Versions: 1.4.10
>Reporter: Stefan Eissing
>Priority: Minor
> Attachments: LuceneIndexConfigObserver-under-load.png
>
>
> The maximum queue length in the {{LuceneIndexConfigObserver}} is hard coded 
> to 5. This is unreasonable short for production systems experiencing heavy 
> load.
> Tests with a patched version resulted in observed queue length of >100 
> entries. This floods warnings into the error log, produces additional load.
> The fix would be to increase the queue max, make it configurable or rely on 
> the system default (which can be configured).



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)