[ 
https://issues.apache.org/jira/browse/MRESOLVER-25?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16028146#comment-16028146
 ] 

Roland Illig commented on MRESOLVER-25:
---------------------------------------

Yes, exactly that.

> Resume support is broken under high concurrency
> -----------------------------------------------
>
>                 Key: MRESOLVER-25
>                 URL: https://issues.apache.org/jira/browse/MRESOLVER-25
>             Project: Maven Resolver
>          Issue Type: Bug
>          Components: resolver
>    Affects Versions: Maven Artifact Resolver 1.0.3
>            Reporter: Roland Illig
>         Attachments: dependency-list.txt, dependency-list-x.zip, 
> wagon-not-threadsafe.zip
>
>
> When building a multi-module project in parallel, Maven updates the local 
> Maven repository without taking into account that other threads or processes 
> might do the same at the same time.
> To reproduce:
> 1.  unpack the attached project
> 2.  {{mvn dependency:list -U -B -T100}}
> See the attached {{dependency-list.txt}} for an example output.
> First thing to notice is that the dependency is downloaded 100 times. This is 
> unexpected, since the Reactor should coordinate all the modules.
> Second thing to notice is that the build fails, since a dependency "cannot be 
> found". This is wrong, since the dependency _can_ be found, it's just not 
> processed properly.
> I suspect the {{legacy.DefaultWagonManager}} to be the cause of this, since 
> the typical error messages are:
> * {{"Downloaded file does not exist: "}}
> * {{"Error copying temporary file to the final destination: "}}
> * {{"*** CHECKSUM FAILED - RETRYING"}}



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

Reply via email to