questions about multiple masters

[Xian Woo <in...@gmail.com>]

Hello, everyone, I've got a question here and I really need your advice.

Background: The Hadoop version is CDH3u1. I've got 8 nodes on my cluster,
one as namenode/jobtracker/master, the others
asdatanode/jtasktracker/regionserver.  I've backed up the namenode by
creating a directory via NFS.

I'm wondering whether I should back up the hbase master or not. Is there any
possibility of data loss if I don't back up the hbase master?

Also, if I back up the hbase master and hope that the backup hbase master
continues to work when the real master not working, should I separate the
hbase master from the hadoop namenode? Here is my
concern: As you know, when the master of the cluster went down, the cluster
would crash(since the  namenode is on the master node). And I suppose the
backup master would  not get to work until the hadoop cluster back to
normal.

I would really appreciate it if you could comment on my question here. Thank
you, guys.

Best wishes,

Xian Wu

RE: questions about multiple masters

[Stuti Awasthi <st...@hcl.com>]

HI Xian,

> I'm wondering whether I should back up the hbase master or not. Is
> there any possibility of data loss if I don't back up the hbase master?

Well if you do not take the backup of Hbase Master, then you will not be able to access your data present in HDFS but since the data is on HDFS so data is not lost. Here I am assuming that your HDFS cluster is up and running.
If you take Backup Hbase Master, then once your Hbase Master is down , after some time Backup Hbase Master will become Hbase Master and you will be able to access your data as before.
In both case if your HDFS cluster is running fine, then your data is safe.

Hope this helps!

Stuti


-----Original Message-----
From: saint.ack@gmail.com [mailto:saint.ack@gmail.com] On Behalf Of Stack
Sent: Thursday, October 20, 2011 9:33 PM
To: user@hbase.apache.org
Subject: Re: questions about multiple masters

On Wed, Oct 19, 2011 at 11:59 PM, Xian Woo <in...@gmail.com> wrote:
> I'm wondering whether I should back up the hbase master or not. Is
> there any possibility of data loss if I don't back up the hbase master?
>

No.  Master does not hold 'state'.  Any state it needs it keeps out in zookeeper.


> Also, if I back up the hbase master and hope that the backup hbase
> master continues to work when the real master not working, should I
> separate the hbase master from the hadoop namenode? Here is my
> concern: As you know, when the master of the cluster went down, the
> cluster would crash(since the  namenode is on the master node). And I
> suppose the backup master would  not get to work until the hadoop
> cluster back to normal.
>

That is right.  HBase won't work when NameNode is down.

St.Ack

::DISCLAIMER::
-----------------------------------------------------------------------------------------------------------------------

The contents of this e-mail and any attachment(s) are confidential and intended for the named recipient(s) only.
It shall not attach any liability on the originator or HCL or its affiliates. Any views or opinions presented in
this email are solely those of the author and may not necessarily reflect the opinions of HCL or its affiliates.
Any form of reproduction, dissemination, copying, disclosure, modification, distribution and / or publication of
this message without the prior written consent of the author of this e-mail is strictly prohibited. If you have
received this email in error please delete it and notify the sender immediately. Before opening any mail and
attachments please check them for viruses and defect.

-----------------------------------------------------------------------------------------------------------------------

Re: questions about multiple masters

[Stack <st...@duboce.net>]

On Wed, Oct 19, 2011 at 11:59 PM, Xian Woo <in...@gmail.com> wrote:
> I'm wondering whether I should back up the hbase master or not. Is there any
> possibility of data loss if I don't back up the hbase master?
>

No.  Master does not hold 'state'.  Any state it needs it keeps out in
zookeeper.


> Also, if I back up the hbase master and hope that the backup hbase master
> continues to work when the real master not working, should I separate the
> hbase master from the hadoop namenode? Here is my
> concern: As you know, when the master of the cluster went down, the cluster
> would crash(since the  namenode is on the master node). And I suppose the
> backup master would  not get to work until the hadoop cluster back to
> normal.
>

That is right.  HBase won't work when NameNode is down.

St.Ack