You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@forrest.apache.org by Brian M Dube <bd...@apache.org> on 2011/03/18 06:15:42 UTC

Re: Locationmap rewrite example, perhaps zone network glitch

On Thu, Feb 24, 2011 at 05:30:48PM +1100, David Crossley wrote:
> Forrestbot wrote:
>  [ snip ]
> >      [java] X [0]                                     /home/forrest/config/forrestbot-trunk/sources/forrest-seed/src/documentation/content/xdocs/rewrite.index	BROKEN: No pipeline matched request: /home/forrest/config/forrestbot-trunk/sources/forrest-seed/src/documentation/content/xdocs/rewrite.index
> >      [java] java.lang.NullPointerException
> >      [java] 	at org.apache.cocoon.environment.AbstractEnvironment.release(AbstractEnvironment.java:563)
>  [ snip ]
> 
> I don't know what causes these occasional failures.
> Could it be a network glitch on the zone server.
> Can Forrest handle it better.
> Need to look into the Cocoon logfiles on our zone server.

Need to look soon after the failure, or we temporarily suspend log
pruning. I hadn't noticed how short-lived the logs are until I went to
check on this and noticed the last failure is already pruned.

-Brian

> -David

Re: Locationmap rewrite example, perhaps zone network glitch

Posted by Brian M Dube <bd...@apache.org>.
On Fri, Mar 18, 2011 at 05:47:17PM +1100, David Crossley wrote:
> Brian M Dube wrote:
> > David Crossley wrote:
> > > Forrestbot wrote:
> > >  [ snip ]
> > > >      [java] X [0]                                     /home/forrest/config/forrestbot-trunk/sources/forrest-seed/src/documentation/content/xdocs/rewrite.index	BROKEN: No pipeline matched request: /home/forrest/config/forrestbot-trunk/sources/forrest-seed/src/documentation/content/xdocs/rewrite.index
> > > >      [java] java.lang.NullPointerException
> > > >      [java] 	at org.apache.cocoon.environment.AbstractEnvironment.release(AbstractEnvironment.java:563)
> > >  [ snip ]
> > > 
> > > I don't know what causes these occasional failures.
> > > Could it be a network glitch on the zone server.
> > > Can Forrest handle it better.
> > > Need to look into the Cocoon logfiles on our zone server.
> > 
> > Need to look soon after the failure, or we temporarily suspend log
> > pruning. I hadn't noticed how short-lived the logs are until I went to
> > check on this and noticed the last failure is already pruned.
> 
> I updated the f.a.o/zone.html doc to better explain the log files.
> Yes the Cocoon ones are of course re-generated on each hourly run.
> Perhaps we could add some post-processing to $CONFIG/run-forrestbot.sh
> to save the Cocoon logs if there is a non-successful build.

I see nothing revealing in the logs from the last failure.

I modified a local seed to try this link rewriting example with a name
that is known not to resolve to IP, and the failure mode is the same
as we are seeing with the zone. An occasional DNS issue?

Also, the NPE quoted above is resolved for me by adding one check for
null to the Cocoon source. Why it is null for our use case and not for
Cocoon's, I haven't investigated.

-Brian

Re: Locationmap rewrite example, perhaps zone network glitch

Posted by David Crossley <cr...@apache.org>.
Brian M Dube wrote:
> David Crossley wrote:
> > Forrestbot wrote:
> >  [ snip ]
> > >      [java] X [0]                                     /home/forrest/config/forrestbot-trunk/sources/forrest-seed/src/documentation/content/xdocs/rewrite.index	BROKEN: No pipeline matched request: /home/forrest/config/forrestbot-trunk/sources/forrest-seed/src/documentation/content/xdocs/rewrite.index
> > >      [java] java.lang.NullPointerException
> > >      [java] 	at org.apache.cocoon.environment.AbstractEnvironment.release(AbstractEnvironment.java:563)
> >  [ snip ]
> > 
> > I don't know what causes these occasional failures.
> > Could it be a network glitch on the zone server.
> > Can Forrest handle it better.
> > Need to look into the Cocoon logfiles on our zone server.
> 
> Need to look soon after the failure, or we temporarily suspend log
> pruning. I hadn't noticed how short-lived the logs are until I went to
> check on this and noticed the last failure is already pruned.

I updated the f.a.o/zone.html doc to better explain the log files.
Yes the Cocoon ones are of course re-generated on each hourly run.
Perhaps we could add some post-processing to $CONFIG/run-forrestbot.sh
to save the Cocoon logs if there is a non-successful build.

-David