Hi,

If we do a bugfix relaese I would like to get the LUCENE-6482 (classloading 
deadlock) in. This is not a new bug, but it is very nasty and affects 
Elasticsearch. Solr is no affected because the Codec API is initialized in 
single-threaded mode. It is not high priority, but this one is easy to backport 
and does not affect any public APIs. There is a workaround available (executing 
Codecs.availableCodecs() on startup of your server), but this is really nasty 
because your JVM just hangs. You don't even see any locked/blocked threads, 
just threads lingering before a method call - without a lock.

Uwe

-----
Uwe Schindler
H.-H.-Meier-Allee 63, D-28213 Bremen
http://www.thetaphi.de
eMail: [email protected]

> -----Original Message-----
> From: Shawn Heisey [mailto:[email protected]]
> Sent: Monday, June 08, 2015 7:06 PM
> To: [email protected]
> Subject: Looks like I broke Solr 5.2.0 - do we need a 5.2.1?
> 
> I broke it.  SOLR-7588 fixes it.  It hasn't been committed yet.
> 
> The dataimport section of the admin UI doesn't work because I apparently
> put coffeescript into the admin UI instead of javascript.  Is this a bad 
> enough
> problem to warrant a bugfix release?
> 
> Thanks,
> Shawn
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected] For additional
> commands, e-mail: [email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to