Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Accumulo, mail # user - Tserver kills themselves from lost Zookeeper locks


Copy link to this message
-
Re: Tserver kills themselves from lost Zookeeper locks
Eric Newton 2013-11-05, 19:01
Limiting the size of the New Generation can avoid long pauses during New
Generation collection:

"-XX:NewSize=1G -XX:MaxNewSize=1G"

And, of course, using the Concurrent collector, tweaked so it kicks in
early enough to avoid stop-the-world pauses:

"-XX:+UseConcMarkSweepGC -XX:CMSInitiatingOccupancyFraction=75"

Both of these are in the example configurations, so you may be using them
already.

-Eric

On Tue, Nov 5, 2013 at 12:44 PM, buttercream <[EMAIL PROTECTED]
> wrote:

> Our machines currently have 8GB of memory, so I'll push that up and see if
> the error occurs again. Also, are there any recommended settings changes to
> the garbage collection that would cause it to be more efficient or at least
> less intrusive?
>
>
>
> --
> View this message in context:
> http://apache-accumulo.1065345.n5.nabble.com/Tserver-kills-themselves-from-lost-Zookeeper-locks-tp6125p6150.html
> Sent from the Users mailing list archive at Nabble.com.
>