You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@archiva.apache.org by "Marc Lustig (JIRA)" <ji...@codehaus.org> on 2009/12/01 10:48:55 UTC

[jira] Created: (MRM-1291) Archiva has periodic load-peaks

Archiva has periodic load-peaks 
--------------------------------

                 Key: MRM-1291
                 URL: http://jira.codehaus.org/browse/MRM-1291
             Project: Archiva
          Issue Type: Bug
          Components: system
    Affects Versions: 1.2.2
         Environment: RHEL4, Oracle RDBMS
            Reporter: Marc Lustig
         Attachments: cpu-day.png

Archiva causes periodically an increased system-load. (see Munin-stats attached)

During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.

I could not identify corresponding messages in the logs.

Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Closed: (MRM-1291) Archiva has periodic load-peaks

Posted by "Brett Porter (JIRA)" <ji...@codehaus.org>.
     [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Brett Porter closed MRM-1291.
-----------------------------

    Resolution: Duplicate
      Assignee: Brett Porter

closing as a duplicate of the various issues solved in 1.3. Please reopen if you continue to experience it.

> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>            Assignee: Brett Porter
>         Attachments: cpu-day.png, cpu-week2.png, screenshot-1.jpg
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Commented: (MRM-1291) Archiva has periodic load-peaks

Posted by "Brett Porter (JIRA)" <ji...@codehaus.org>.
    [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=206999#action_206999 ] 

Brett Porter commented on MRM-1291:
-----------------------------------

the release candidate is now being voted on (it is now labeled 1.3). If you get a chance to, please test it and let us know if that helps.

> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>         Attachments: cpu-day.png, cpu-week2.png, screenshot-1.jpg
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Commented: (MRM-1291) Archiva has periodic load-peaks

Posted by "Marc Lustig (JIRA)" <ji...@codehaus.org>.
    [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=201533#action_201533 ] 

Marc Lustig commented on MRM-1291:
----------------------------------

According to Brett everything will be solved in 1.2.3
Looking forward to test the release version....

> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>         Attachments: cpu-day.png, cpu-week2.png, screenshot-1.jpg
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Issue Comment Edited: (MRM-1291) Archiva has periodic load-peaks

Posted by "Marc Lustig (JIRA)" <ji...@codehaus.org>.
    [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=200559#action_200559 ] 

Marc Lustig edited comment on MRM-1291 at 12/4/09 3:23 AM:
-----------------------------------------------------------

I want to provide you a third chart (cpu-week2.png).
The filled area shows the "blown-up" load on the left half.
The second half shows the normal load.
All seems to be caused by the periodic scan processes.


      was (Author: axismundi):
    I want to provide you a third chart.
The filled area shows the "blown-up" load on the left half.
The second half shows the normal load.
All seems to be caused by the periodic scan processes.

  
> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>         Attachments: cpu-day.png, cpu-week2.png, screenshot-1.jpg
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Commented: (MRM-1291) Archiva has periodic load-peaks

Posted by "Brett Porter (JIRA)" <ji...@codehaus.org>.
    [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=200521#action_200521 ] 

Brett Porter commented on MRM-1291:
-----------------------------------

did you have any large change in artifacts / rebuild the index in that period?

I'd like to encourage you to try 1.2.3-SNAPSHOT or the release - the hourly peaks will certainly be reduced. If it seems managable, we can close this out.

> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>         Attachments: cpu-day.png, screenshot-1.jpg
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Commented: (MRM-1291) Archiva has periodic load-peaks

Posted by "Marc Lustig (JIRA)" <ji...@codehaus.org>.
    [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=200153#action_200153 ] 

Marc Lustig commented on MRM-1291:
----------------------------------

This pretty much sounds like a reasonable explanation - unfortunately it may not be the (only) reason.

I forgot to mention that this "periodic peaks" load-pattern occurs only from time to time.
I. e. when the pattern happens, the peaks occur periodically. But after restarting Archiva, the pattern does not occur anymore for a couple of days or weeks.

I can be sure about this because we have notifiers installed that warn when a certain HTTP-request on Archiva exceeds a certain threshold value.
We have not seen alarms between 3rd of november and 30th of november.
On the 30th out of a sudden the periodic load raised tremendously.

I doubt the RDBMS being slower could have an impact on the load.



> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>         Attachments: cpu-day.png
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Issue Comment Edited: (MRM-1291) Archiva has periodic load-peaks

Posted by "Marc Lustig (JIRA)" <ji...@codehaus.org>.
    [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=200153#action_200153 ] 

Marc Lustig edited comment on MRM-1291 at 12/2/09 2:43 AM:
-----------------------------------------------------------


I forgot to mention that this "periodic peaks" load-pattern occurs only from time to time.
I. e. when the pattern happens, the peaks occur periodically. But after restarting Archiva, the pattern does not occur anymore for a couple of days or weeks.

I can be sure about this because we have notifiers installed that warn when a certain HTTP-request on Archiva exceeds a certain threshold value.
We have not seen alarms between 3rd of november and 30th of november.
On the 30th out of a sudden the periodic load raised tremendously.

I doubt the RDBMS being slower could have an impact on the load.



      was (Author: axismundi):
    This pretty much sounds like a reasonable explanation - unfortunately it may not be the (only) reason.

I forgot to mention that this "periodic peaks" load-pattern occurs only from time to time.
I. e. when the pattern happens, the peaks occur periodically. But after restarting Archiva, the pattern does not occur anymore for a couple of days or weeks.

I can be sure about this because we have notifiers installed that warn when a certain HTTP-request on Archiva exceeds a certain threshold value.
We have not seen alarms between 3rd of november and 30th of november.
On the 30th out of a sudden the periodic load raised tremendously.

I doubt the RDBMS being slower could have an impact on the load.


  
> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>         Attachments: cpu-day.png, screenshot-1.jpg
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Updated: (MRM-1291) Archiva has periodic load-peaks

Posted by "Marc Lustig (JIRA)" <ji...@codehaus.org>.
     [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Marc Lustig updated MRM-1291:
-----------------------------

    Attachment: screenshot-1.jpg

Please have a look at the second stat-chart by Munin.
Unfortunately we miss the data between week 46 and 48.

But when you look at the filled area in week 45, you can see very the apparent load-difference between normal scanning peaks and "overblown" scanning-peaks.

By the way, we do not have any other software running on that server.

> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>         Attachments: cpu-day.png, screenshot-1.jpg
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] Updated: (MRM-1291) Archiva has periodic load-peaks

Posted by "Marc Lustig (JIRA)" <ji...@codehaus.org>.
     [ http://jira.codehaus.org/browse/MRM-1291?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Marc Lustig updated MRM-1291:
-----------------------------

    Attachment: cpu-week2.png

I want to provide you a third chart.
The filled area shows the "blown-up" load on the left half.
The second half shows the normal load.
All seems to be caused by the periodic scan processes.


> Archiva has periodic load-peaks 
> --------------------------------
>
>                 Key: MRM-1291
>                 URL: http://jira.codehaus.org/browse/MRM-1291
>             Project: Archiva
>          Issue Type: Bug
>          Components: system
>    Affects Versions: 1.2.2
>         Environment: RHEL4, Oracle RDBMS
>            Reporter: Marc Lustig
>         Attachments: cpu-day.png, cpu-week2.png, screenshot-1.jpg
>
>
> Archiva causes periodically an increased system-load. (see Munin-stats attached)
> During these peaks calling the homepage takes around 6 secs, whereas normally it is around 1 sec.
> I could not identify corresponding messages in the logs.
> Could somebody else observe this?

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira