hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Vamshi Krishna <vamshi2...@gmail.com>
Subject Re: HRegionserver daemon is not running on region server node
Date Mon, 28 Nov 2011 06:58:29 GMT
Hi,
1)No, hbase is running as same user i.e hduser, in all machines i have
hbase in hduser account only.

2)Yes, iset HBASE_HOME environmental variable in all machines under hduser
, in the file /.bashrc  like
  export HBASE_HOME=/home/hduser/Documents/HBASE_SOFTWRAE/hbase-0.90.4
And, in all the machines, hbase is located at the same place, i.e if i  run
command   echo $HBASE_HOME in all machines separatley ,under hduser
account, three machines are printing
/home/hduser/Documents/HBASE_SOFTWRAE/hbase-0.90.4    , So, i hope
HBASE_HOME is set in all machines, right?

3) i didn't understand " If a new bash env is created, will that env get
inherited?" statement. I didn't create any new bash env.

4) i changed all the machines to wired network only. Now all machines are
connected with wired connection.
5)hbase is  running fine in standalone mode.

 Hadoop is running all machines perfectly, i ran some example(word count) ,
its running fine on all machines, in distributed mode.
Hbase is perfectly running on master node ONLY.
This is my present situation, i am feeling like just struck with this
problem, please help.


On Mon, Nov 28, 2011 at 2:04 AM, Suraj Varma <svarma.ng@gmail.com> wrote:

> So - first of all:
> > hbase-regionserver2: bash: line 0: cd: {HBASE_HOME}/bin/..:
> > No such file or
> > directory
>
> This indicates that the environment of the process running
> hbase-regionserver2 is not set correctly. It is not able to find the
> directory.
>
> Is hbase running as a different user?
>
> I think your first step should be to get the above "No such file or
> directory" errors resolved. Did you set HBASE_HOME in the  environment
> manually? If a new bash env is created, will that env get inherited?
>
> So - start simple steps first - get hbase running stand alone, then
> move to pseudo / fully distributed.
>
> By introducing variables like wired/wireless etc in your network, you
> are introducing uncertainities which will bite you on an ongoing basis
> (wireless going down, increased latencies, etc, etc). So - the cluster
> is unlikely to be stable with such a setup.
>
> I would recommend going with a fully wired setup, if your goal is to
> have a stable hbase cluster. If it is a "at home test cluster", then
> that's fine - but be prepared for frequent cluster reboots. :)
>
> --Suraj
>
>
> On Sat, Nov 26, 2011 at 11:20 PM, Vamshi Krishna <vamshi2105@gmail.com>
> wrote:
> > Hi suraj,
> > Thank you,
> >  i am using ubuntu-10.04 OS. actually i am using 3 machines out of which
> > two are desktops connected with wired connection(10.0.1.54 and
> 10.0.1.97),
> > other is laptop(10.0.1.55) connected with wireless router. i didnot
> modify
> > any of such interface settings in hbase-site.xml file.
> > Would this sort of connecting nodes (i.e using both wired and wireless
> > connections) create any such problem as i mentioned? If so, atleast two
> > desktop machines are connected with same wired connection, then also  it
> > threw an error, like
> >
> > hbase-regionserver2: bash: line 0: cd: {HBASE_HOME}/bin/..: No such file
> or
> > directory
> > hbase-regionserver2: bash: {HBASE_HOME}/bin/hbase-daemon.sh: No such
> > file or directory
> >
> > i will change hbase-site.xml as you have suggested and let you know what
> is
> > happening. But before that answer my above query please.
> >
> >  One more thing is, i am running  region servers on the datanodes and
> > master only, then i hope it would become co-located, right?
> >
> > Thank you
> > On Sun, Nov 27, 2011 at 2:05 AM, Suraj Varma <svarma.ng@gmail.com>
> wrote:
> >
> >> Vamshi:
> >> What OS are you trying this on? Is it Linux / Windows?
> >>
> >> You can use the *.dns.interface configurations to use the specific
> >> network interface that you want.
> >>
> >> That is, set the following configurations in your hbase-site.xml on
> >> all hbase nodes
> >> hbase.zookeeper.dns.interface
> >> hbase.regionserver.dns.interface
> >> hbase.master.dns.interface
> >>
> >> You can set it to lo or eth0 or eth1 per your environment.
> >>
> >> Restart all nodes after this and you should see them bind to the
> >> specified interface.
> >> You can do the same for hadoop by setting dfs.datanode.dns.interface
> >> in hdfs-site.xml for hadoop nodes.
> >>
> >> On Windows cygwin, I've found setting that it to "localhost" works.
> >>
> >> Your setup should work for test environments ... for production, the
> >> standard setup would be to co-locate region servers and data nodes to
> >> get data locality.
> >> --Suraj
> >>
> >> On Thu, Nov 24, 2011 at 10:51 PM, Vamshi Krishna <vamshi2105@gmail.com>
> >> wrote:
> >> > Hi, i checked  by executing netstat -anp, i found
> >> > tcp        0      0 127.0.1.1:60000         0.0.0.0:*
> >> > LISTEN      6814/java
> >> > .
> >> > ***some other ports**
> >> > tcp        0      0 0.0.0.0:60010           0.0.0.0:*
> >> > LISTEN      6814/java .
> >> > ***some other ports**
> >> > .
> >> > tcp        0      0 10.0.1.54:44095         10.0.1.54:9000
> >> > ESTABLISHED 6814/java
> >> > ***some other ports**
> >> > tcp        0      0 10.0.1.54:55204         10.0.1.54:2222
> >> > ESTABLISHED 6814/java
> >> > ***some other ports**
> >> > ***some other ports**
> >> >
> >> > the port of HMaster i found when i ran jps is 6814. But the same port
> i
> >> > found in different places bound to different IPs (i.e some time
> >> 127.0.1.1 ,
> >> > and some times 10.0.1.54). Same is the case with all other daemons
> which
> >> > are running after i ran bin/start-hbase.sh
> >> > Does it cause any abnormal behavior if i have like above. All hadoop
> >> > daemons are runnong on other machines, but hbase daemon i.e
> HRegionServer
> >> > is not running on other machines. HMaster, HQuorumPeer, HRegionServer
> are
> >> > running on master node.
> >> > My /etc/hosts file has content like
> >> >
> >> > 127.0.0.1       localhost
> >> > 127.0.1.1       vamshikrishna-desktop
> >> > 10.0.1.54       hbase-master      hadoop-namenode
> >> > 10.0.1.55       hbase-regionserver1     hadoop-datanode1
> >> >
> >> > Still what more modification i need to do to bind the hostnames  to
> IPs?
> >> >
> >> >
> >> > Please some body help..
> >> >
> >> >
> >> > On Mon, Nov 21, 2011 at 7:42 PM, Joey Echeverria <joey@cloudera.com>
> >> wrote:
> >> >
> >> >> I suspect that the daemon's are binding to 127.0.0.1 instead of
> >> >> 10.0.1.54 on the master. You can check this by running netstat -anp.
> I
> >> >> would recommend either configuring the /etc/hosts to bind the
> >> >> vamshikrishna-desktop and vamshi-laptop hostnames to the 10.0.1.x
> >> >> addresses.
> >> >>
> >> >> -Joey
> >> >>
> >> >> On Thu, Nov 17, 2011 at 1:40 AM, Vamshi Krishna <
> vamshi2105@gmail.com>
> >> >> wrote:
> >> >> > hi
> >> >> > i am working with 2 node hbase cluster as shown below
> >> >> > On  node1 (10.0.1.54) : master node, region server, hadoop
> namenode,
> >> >> hadoop
> >> >> > datanode
> >> >> > on node2 (10.0.1.55): region server, hadoop datanode.
> >> >> > When i start both hadoop then hbase, all daemons are running
> properly
> >> on
> >> >> > masternode i.e node1,
> >> >> > 2404 NameNode
> >> >> > 3657
> >> >> > 3007 TaskTracker
> >> >> > 2848 JobTracker
> >> >> > 3522 HRegionServer
> >> >> > 3848 Main
> >> >> > 3292 HQuorumPeer
> >> >> > 2769 SecondaryNameNode
> >> >> > 3345 HMaster
> >> >> > 2575 DataNode
> >> >> > 4768 Jps
> >> >> >  but on node2, only TaskTracker and Datanode daemons are running,
> the
> >> >> > HRegionServer daemon is not running.
> >> >> > I provide some files which may be helpful in helping me. i have
my
> >> >> > /etc/hosts file of node1 (vamshikrishna-desktop) has
> >> >> > # /etc/hosts (for master AND slave)
> >> >> >
> >> >> > 127.0.0.1       localhost
> >> >> > 127.0.1.1       vamshikrishna-desktop
> >> >> > 10.0.1.54 hbase-master  hadoop-namenode
> >> >> > 10.0.1.55 hbase-regionserver1   hadoop-datanode1
> >> >> >
> >> >> > and file {HBASE_HOME}/conf/regionservers has
> >> >> >
> >> >> > hbase-regionserver1
> >> >> > hbase-master
> >> >> >
> >> >> > i have my /etc/hosts file of node2 (vamshikrishna-laptop) has
> >> >> > 127.0.0.1    localhost
> >> >> > 127.0.0.1    vamshi-laptop
> >> >> > 10.0.1.54 hbase-master  hadoop-namenode
> >> >> > 10.0.1.55 hbase-regionserver1 hadoop-datanode1
> >> >> >
> >> >> > and file {HBASE_HOME}/conf/regionservers has
> >> >> >  localhost
> >> >> >
> >> >> > Please can anybody help me in finding out what is the reason for
> >> >> > Regionserver not running on node2..?  And moreover when i run
> >> >> > start-hbase.sh , it will display error like
> >> >> > hbase-regionserver1: bash: line 0: cd: {HBASE_HOME}/bin/..: No
such
> >> file
> >> >> or
> >> >> > directory
> >> >> > hbase-regionserver1: bash: {HBASE_HOME}/bin/hbase-daemon.sh: No
> such
> >> file
> >> >> > or directory  , but i could find out
> {HBASE_HOME}/bin/hbase-daemon.sh
> >> >> > clearly.. i don't know what went wrong..!
> >> >> >  please can any body help?
> >> >> >
> >> >> > --
> >> >> > *Regards*
> >> >> > *
> >> >> > Vamshi Krishna
> >> >> > *
> >> >> >
> >> >>
> >> >>
> >> >>
> >> >> --
> >> >> Joseph Echeverria
> >> >> Cloudera, Inc.
> >> >> 443.305.9434
> >> >>
> >> >
> >> >
> >> >
> >> > --
> >> > *Regards*
> >> > *
> >> > Vamshi Krishna
> >> > *
> >> >
> >>
> >
> >
> >
> > --
> > *Regards*
> > *
> > Vamshi Krishna
> > *
> >
>



-- 
*Regards*
*
Vamshi Krishna
*

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