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

Switch to Threaded View
Hadoop, mail # user - monit? daemontools? jsvc? something else?


Copy link to this message
-
Re: monit? daemontools? jsvc? something else?
Otis Gospodnetic 2011-01-06, 08:34
So Allen, what do you use to monitor those processes/nodes?

Thanks,
Otis
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch
Hadoop ecosystem search :: http://search-hadoop.com/

----- Original Message ----
> From: Allen Wittenauer <[EMAIL PROTECTED]>
> To: "<[EMAIL PROTECTED]>" <[EMAIL PROTECTED]>
> Sent: Wed, January 5, 2011 11:54:22 PM
> Subject: Re: monit? daemontools? jsvc? something else?
>
>
> On Jan 5, 2011, at 7:57 PM, Lance Norskog wrote:
>
> > Isn't this what  Ganglia is for?
> >
>
>     No.
>
>      Ganglia does metrics, not monitoring.
>
>
> > On 1/5/11, Allen  Wittenauer <[EMAIL PROTECTED]>  wrote:
> >>
> >> On Jan 4, 2011, at 10:29 PM, Otis Gospodnetic  wrote:
> >>
> >>> Ah, more manual work! :(
> >>>
> >>> You guys never have JVM die "just because"? I just had a DN's  JVM die the
> >>> other day "just because and with no obvious  cause".  Restarting it
brought
> >>> it
> >>> back to  life, everything recovered smoothly.  Had some automated tool  
>done
> >>> the
> >>> restart for me, I'd be even  happier.
> >>
> >>     In the case of Hadoop,  no.  There has usually been at least a core
>dump,
> >> message in  syslog, message in datanode log, etc, etc.   [You *do* have  
>cores
> >> enabled, right?]
> >>
> >>      We also have in place a monitor that checks the # of active nodes.  If  
>it
> >> falls below a certain percentage, then we get alerted and check  on them en
> >> masse.   Worrying about one or two nodes going down  probably means you
need
> >> more nodes. :D
> >>
> >>
> >
> >
> > --
> > Lance Norskog
> > [EMAIL PROTECTED]
>
>