On Fri, Jun 7, 2013 at 9:10 AM, Nick Niemeyer <[email protected]>wrote:
> Regarding TotalLoad, what would be normal operating tolerances per node > for this metric? When should one become concerned? Thanks again to > everyone participating in this community. :) > > Why do you want to be concered :) I have not seen many issues related to high TotalLoad. This is mainly useful in terms of understanding how many concurrent jobs/file accesses are happening and how busy datanodes are. When you are debugging issues where cluster slow down due to overload, or correlating a run of big jobs, this is useful. Knowing what it represent, you would find many other uses as well. > From: Suresh Srinivas <[email protected]> > Reply-To: "[email protected]" <[email protected]> > Date: Thursday, June 6, 2013 4:14 PM > To: "[email protected]" <[email protected]> > Subject: Re: Please explain FSNamesystemState TotalLoad > > It is the total number of transceivers (readers and writers) reported > by all the datanodes. Datanode reports this count in periodic heartbeat to > the namenode. > > > On Thu, Jun 6, 2013 at 1:48 PM, Nick Niemeyer <[email protected]>wrote: > >> Can someone please explain what TotalLoad represents below? Thanks >> for your response in advance! >> >> Version: hadoop-0.20-namenode-0.20.2+923.197-1 >> >> Example pulled from the output of via the name node: >> # curl -i http://localhost:50070/jmx >> >> { >> "name" : "hadoop:service=NameNode,name=FSNamesystemState", >> "modelerType" : "org.apache.hadoop.hdfs.server.namenode.FSNamesystem", >> "CapacityTotal" : #, >> "CapacityUsed" : #, >> "CapacityRemaining" : #, >> * "TotalLoad" : #,* >> "BlocksTotal" : #, >> "FilesTotal" : #, >> "PendingReplicationBlocks" : 0, >> "UnderReplicatedBlocks" : 0, >> "ScheduledReplicationBlocks" : 0, >> "FSState" : "Operational" >> } >> >> >> Thanks, >> Nick >> > > > > -- > http://hortonworks.com/download/ > -- http://hortonworks.com/download/
