hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Derek Wollenstein <de...@klout.com>
Subject Re: Help Tuning HBase GC Settings
Date Wed, 07 Dec 2011 18:17:05 GMT
St.Ack --
    We are definitely running with a high handler count -- I've pastebin'd
my config to http://pastebin.com/i6PXixYF.  The boxes are seeing moderate,
but not especially high traffic (topping out around 1000qps).  These
problems do not seem to occur at the same time as request peaks, nor at the
same time as compactions (or at least not regularly;  They do occur during
compactions, but also when no compactions are running)

--Derek

On Wed, Dec 7, 2011 at 8:56 AM, Stack <stack@duboce.net> wrote:

> On Tue, Dec 6, 2011 at 6:30 PM, Derek Wollenstein <derek@klout.com> wrote:
> > Jacques --
> >    The main problem we really see is that after "a while"(yes, this is
> > unspecific) we start seeing "random" (e.g. uncorrelated with the key
> being
> > retrieved, but correlated with time) timeouts (>500ms) retrieving small
> (<
> > 200 byte) rows from hbase.
>
> What percentage of your reads do this?  All reads around the time of a
> YGC flurry.
>
> (If you let server run, does it ever move on past many YGCs or does it
> stay there once it hits the condition?)
>
> How many handlers are you running with?  Pastebin your configuration
> too?  Whats the traffic like on these boxes?  How many
> requests/second?  Is this happening during a traffic peak?
>
> St.Ack
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message