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

Switch to Plain View
Accumulo, mail # dev - tserver.readahead.concurrent.max set to 64 but only running 32 at most.


+
David Medinets 2013-05-15, 20:44
+
Eric Newton 2013-05-16, 17:25
+
David Medinets 2013-05-16, 18:14
Copy link to this message
-
Re: tserver.readahead.concurrent.max set to 64 but only running 32 at most.
Keith Turner 2013-05-17, 17:10
On Thu, May 16, 2013 at 2:14 PM, David Medinets <[EMAIL PROTECTED]>wrote:

> Thanks. I double-checked our system by running config in the shell (as
> opposed looking in the configuration file). The config command reported a
> 'system' override of 32. So that explains the limit that I saw. I'm not
> sure where that 'system' override came from.
>

Thats coming from a system config set in the shell, which is stored in
zookeeper.
>
>
> On Thu, May 16, 2013 at 1:25 PM, Eric Newton <[EMAIL PROTECTED]>
> wrote:
>
> > tserver.scan.files.open.max
> >
> >
> >
> > On Wed, May 15, 2013 at 4:44 PM, David Medinets <
> [EMAIL PROTECTED]
> > >wrote:
> >
> > > When tserver.readahead.concurrent.max was set to its default setting,
> the
> > > load on our tservers was fairly low. So I changed accumulo-site.xml so
> > the
> > > max is 64. (Go for the gusto!) Then restarted the tservers to ensure
> the
> > > change was picked up.
> > >
> > > New queries are now running 32 concurrent scans on each server with
> > several
> > > scans waiting in the queue. Is there some other accumulo parameter
> that I
> > > need to set to get more than 32 concurrent scans? Is something else
> > > affecting the scans?
> > >
> >
>