You are viewing a plain text version of this content. The canonical link for it is here.

Posted to common-user@hadoop.apache.org by prasenjit mukherjee <pm...@quattrowireless.com> on 2010/02/15 03:57:14 UTC

launching time for high-cpu-instance hadoop clusters

I am using hadoop's in-built ec2-scripts to launch hadoop clusters in
ec2.  For lower instance-types ( upto m2.2xlarge ) the launch scripts
work fine. But when I move to either m2.4xlarge or c1.xlarge
instances, the same scripts seem to get stuck. For c1.xlarge it gives
connection-time out and for m2.4xlarge it just didnt seem to proceed
further after launching the master.   Since I am using the exact same
AMI ( because m2.2xlarge runs fine with 64 bit setup ) , I wasn't sure
whether the problem is in AWS's ec2 or I need to do tweak something on
my side.

Thanks,
-Prasen

Re: launching time for high-cpu-instance hadoop clusters

Posted by prasenjit mukherjee <pm...@quattrowireless.com>.

Got it solved. I was probably using an older version of hadoop-ec2.sh
which was pointing to an incorrect kernel ( i.e. aki-9800e5f1 ) which
was causing the problem. I commented that out ( in hadoop-ec2-env.sh )
and it is working fine now.

-Thanks,
Prasen

On Mon, Feb 15, 2010 at 8:27 AM, prasenjit mukherjee
<pm...@quattrowireless.com> wrote:
> I am using hadoop's in-built ec2-scripts to launch hadoop clusters in
> ec2.  For lower instance-types ( upto m2.2xlarge ) the launch scripts
> work fine. But when I move to either m2.4xlarge or c1.xlarge
> instances, the same scripts seem to get stuck. For c1.xlarge it gives
> connection-time out and for m2.4xlarge it just didnt seem to proceed
> further after launching the master.   Since I am using the exact same
> AMI ( because m2.2xlarge runs fine with 64 bit setup ) , I wasn't sure
> whether the problem is in AWS's ec2 or I need to do tweak something on
> my side.
>
> Thanks,
> -Prasen
>