As far as I can tell, it is specifically Maven which knocks our server 
out. I don't know if it is notoriously a bad http client or what, but it 
seems that when one or more people are hitting the maven repo we end up 
hitting our max connection limit in apache (200 or so connections) and 
the old connections take forever to die - in fact they usually get 
locked in close_wait status and never die. I suppose it is just as 
likely (if not more so) that our out of date apache is to blame.

If others (osgeo, opengeo) also find running a maven repo to be heavy on 
resources, I wonder if there is a way we could do some kind of 
mirroring/load-balancing/fall-back setup between the three of us to get 
optimal performance and uptime while sharing the load? I'm not that 
familiar with how maven works but it seems like something like this must 
be possible...

Chris

Daniele Romagnoli wrote:
> Hi lists,
> first of all, sorry for cross posting.
>
> Could we select an unique commonly used remote repo where to deploy 
> maven artifacts?
> Refractions 2 still go down, sometimes.
> Therefore, in the past, someone proposed to set a new repo (OSGeo) 
> where to deploy artifacts.
> However, as Justin reported some months ago, the OSGeo server bombs 
> out quite often and therefore the GeoServer nightly build is deploying 
> artifacts on a new additional maven repository, hosted by OpenGeo.
> However, it seems that the geotools root pom hosted at the OSGgeo repo 
> seems pointing against the refractions repo. Therefore some recent 
> updates seem missing. 
> (http://download.osgeo.org/webdav/geotools/org/geotools/geotools/2.5-SNAPSHOT/geotools-2.5-SNAPSHOT.pom)
>
> Can we move to a single unified repo? Any news about the OSGeo 
> deploying issues? (http://trac.osgeo.org/osgeo/ticket/344)
> In case this has been fixed, can you update the automatic build 
> scripts to work against the OSGeo one (in order to updates artifacts 
> and poms)?
> In case OSGeo still have problems, is the OpenGeo repo the most 
> valuable one to be used? (In that case, can you provide me the 
> distribution management tag (URL) to be used to deploy our imageio-ext 
> artifacts on it?)
>
> I guess we need to update these things since there is a bit of 
> confusion about which project relies on which repo and which repo is 
> up to date. 
> Please, let me know.
>
> Best Regards,
> Daniele
>
> -- 
> -------------------------------------------------------
> Eng. Daniele Romagnoli
> Software Engineer
>
> GeoSolutions S.A.S.
> Via Carignoni 51
> 55041 Camaiore (LU)
> Italy
>
> phone: +39 0584983027
> fax:     +39 0584983027
> mob:   +39 328 0559267
>
>
> http://www.geo-solutions.it
>
> -------------------------------------------------------
>
> ------------------------------------------------------------------------
>
> ------------------------------------------------------------------------------
> Register Now & Save for Velocity, the Web Performance & Operations 
> Conference from O'Reilly Media. Velocity features a full day of 
> expert-led, hands-on workshops and two days of sessions from industry 
> leaders in dedicated Performance & Operations tracks. Use code vel09scf 
> and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
> ------------------------------------------------------------------------
>
> _______________________________________________
> Geoserver-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/geoserver-devel
>   


------------------------------------------------------------------------------
Register Now & Save for Velocity, the Web Performance & Operations 
Conference from O'Reilly Media. Velocity features a full day of 
expert-led, hands-on workshops and two days of sessions from industry 
leaders in dedicated Performance & Operations tracks. Use code vel09scf 
and Save an extra 15% before 5/3. http://p.sf.net/sfu/velocityconf
_______________________________________________
Geoserver-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geoserver-devel

Reply via email to