That's *very* odd. A couple of things to check -
Does this happen only with remote artifacts? With remote artifacts from a
specific remote repo?
Are your remote connections going trough a proxy of some kind or a frontend
Apache?
Do you run Artifactory using a custom JDK?
Thanks,
Yoav
On Fri, Jun 18, 2010 at 3:09 PM, Yves Schumann <[email protected]>wrote:
>
> Hi
>
> We have a strange problem with our build system and currently I have no
> idea
> where to start to fix it. We're using Hudson and Artifactory and some jobs
> fail with messages like the following:
>
> [INFO]
> ------------------------------------------------------------------------
> [ERROR] BUILD FAILURE
> [INFO]
> ------------------------------------------------------------------------
> [INFO] Compilation failure
>
> error: error reading
>
> D:\hudson\jobs\r-2.0__Fullbuild_spezial\workspace\.repository\commons-modeler\commons-modeler\2.0.1\commons-modeler-2.0.1.jar;
> error in opening zip file
> error: error reading
>
> D:\hudson\jobs\r-2.0__Fullbuild_spezial\workspace\.repository\commons-digester\commons-digester\2.0\commons-digester-2.0.jar;
> error in opening zip file
> error: error reading
>
> D:\hudson\jobs\r-2.0__Fullbuild_spezial\workspace\.repository\commons-logging\commons-logging-api\1.0.4\commons-logging-api-1.0.4.jar;
> error in opening zip file
> error: error reading
>
> D:\hudson\jobs\r-2.0__Fullbuild_spezial\workspace\.repository\org\asteriskjava\asterisk-java\0.3.1\asterisk-java-0.3.1.jar;
> error in opening zip file
>
>
> If I open such an archive there is a file in the archive with the same name
> as the archive but without the extension. So for the example above there is
> the jar "asterisk-java-0.3.1.jar" which contains a file
> "asterisk-java-0.3.1". This file inside the archive is another (!) archive
> which could be opened and which contains the real content. To make it
> clear:
>
> asterisk-java-0.3.1.jar
> + asterisk-java-0.3.1
> ++ META-INF
> ++ org
>
> So it looks like that the downloaded archives contains the archive again.
> If
> I download the artifact by hand from the Artifactory web frontend it looks
> like described above, an archive inside an archive. Strange also: These
> corrupt artifacts have mostly (but not always) no pom file in Artifactory!?
> There is only the corrupt jar. Artifactory tells me the URL [1] as the
> source of the jar. The direct download from this URL is ok, the content of
> the jar is as expected. So it seems to me that Artifactory has done
> something nasty with the archive!?
>
> If I wipe out the entry on Artifactory using the web frontend, delete the
> corrupt archive on the local repository and start the build on hudson, the
> artifact must be downloaded again which leads to success in most cases. It
> seems to me that this behaviour happens if Artifactory has to download a
> lot
> of new artifacts.
>
> We use Artifactory 2.1.3 (r9204) and for testing purposes 2.2.3 (r10588),
> same behaviour on both installations. The machines are virtual machines
> running on VMware ESXi4 under the linux distribution "eisfair" [2]. There
> are no other services running, just Artifactory.
>
> Any ideas where to dig for the problem are appreciated!
>
>
> [1]
>
> http://repo1.maven.org/maven2/org/asteriskjava/asterisk-java/0.3.1/asterisk-java-0.3.1.jar
> [2] http://eisfair.org
>
> Kind regards,
>
> Yves
> --
> View this message in context:
> http://forums.jfrog.org/Artifactory-delivers-corrupt-artifacts-tp5195007p5195007.html
> Sent from the Artifactory - Users mailing list archive at Nabble.com.
>
>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit. See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Artifactory-users mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/artifactory-users
>
------------------------------------------------------------------------------
ThinkGeek and WIRED's GeekDad team up for the Ultimate
GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
lucky parental unit. See the prize list and enter to win:
http://p.sf.net/sfu/thinkgeek-promo
_______________________________________________
Artifactory-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/artifactory-users
