StarCluster - Mailing List Archive

Re: EBS vol on starcluster

From: Stuart Young <no email>
Date: Fri, 04 Feb 2011 00:34:35 -0500

Hi Archie,

I came across the same problem and it turned out to be due to the volume
not being partitioned although it was formatted correctly.

This documentation on Stardev covers manually partitioning and formatting:
http://web.mit.edu/stardev/cluster/docs/create_volume_manually.html#partitioning-and-formatting-the-new-volume

Try the following steps to create the volume on 'myInstance', a separate
(e.g., non-Starcluster) EC2 instance then detach it ready for
incorporation into your new Starcluster instance. (NB: Expected output
is indented - you should see something like it when you run the commands.)

1. ON myInstance, CREATE A VOLUME (using your volume name as an example):

ec2-create-volume --availability-zone us-east-1a --size 40

     VOLUME vol-521d803a 40 us-east-1a
creating 2011-01-05T15:34:28+0000

ec2-attach-volume vol-521d803a -i i-b42f3fd9 -d /dev/sdz

     ATTACHMENT vol-521d803a i-b42f3fd9 /dev/sdz
attaching 2011-01-05T15:36:28+0000

ec2-describe-volumes


2. PARTITION THE VOLUME WITH ONE LINUX ext2 PARTITION USING THE WHOLE VOLUME
(NB: ext2 is the format of starcluster AMI partitions but in theory ext3
is fine)

echo ",,L" | sfdisk -L /dev/sdz

     Checking that no-one is using this disk right now ...
     OK

     Disk /dev/sdz: 5221 cylinders, 255 heads, 63 sectors/track
     Old situation:
     Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting
from 0

        Device Boot Start End #cyls #blocks Id System
     /dev/sdz1 0+ 5220 5221- 41937682 83 Linux
     /dev/sdz2 0 - 0 0 0 Empty
     /dev/sdz3 0 - 0 0 0 Empty
     /dev/sdz4 0 - 0 0 0 Empty
     New situation:
     Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting
from 0

        Device Boot Start End #cyls #blocks Id System
     /dev/sdz1 0+ 5220 5221- 41937682 83 Linux
     /dev/sdz2 0 - 0 0 0 Empty
     /dev/sdz3 0 - 0 0 0 Empty
     /dev/sdz4 0 - 0 0 0 Empty
     Warning: no primary partition is marked bootable (active)
     This does not matter for LILO, but the DOS MBR will not boot this disk.
     Successfully wrote the new partition table

     Re-reading the partition table ...

     If you created or changed a DOS partition, /dev/foo7, say, then use
dd(1)
     to zero the first 512 bytes: dd if=/dev/zero of=/dev/foo7 bs=512
count=1


3. FORMAT THE NEWLY CREATED PARTITION (NB: ***/dev/sdz1*** ):

mkfs.ext3 /dev/sdz1

     mke2fs 1.39 (29-May-2006)
     Filesystem label=
     OS type: Linux
     Block size=4096 (log=2)
     Fragment size=4096 (log=2)
     5242880 inodes, 10484420 blocks
     524221 blocks (5.00%) reserved for the super user
     First data block=0
     Maximum filesystem blocks=4294967296
     320 block groups
     32768 blocks per group, 32768 fragments per group
     16384 inodes per group
     Superblock backups stored on blocks:
             32768, 98304, 163840, 229376, 294912, 819200, 884736,
1605632, 2654208,
             4096000, 7962624

     Writing inode tables: done
     Creating journal (32768 blocks): done
     Writing superblocks and filesystem accounting information: done

     This filesystem will be automatically checked every 36 mounts or
     180 days, whichever comes first. Use tune2fs -c or -i to override.


4. MOUNT THE NEWLY CREATED PARTITION ON myInstance (NB: ***/dev/sdz1*** ):
mount -t ext2 /dev/sdz1 /scvol

5. COPY OVER DATA FROM /data TO /scvol
cp -rp /data/* /scvol

6. UNMOUNT DEVICE AND DETACH VOLUME FROM myInstance
umount /dev/sdz1
ec2-detach-volume vol-521d803a


7. ADD [volume ...] SECTION TO STARCLUSTER CONFIG
(You can call it anything you like but I used 'data' mounting to the
folder '/data'.)

[volume data]
DEVICE=/dev/sdz
MOUNT_PATH=/data
PARTITION=1
VOLUME_ID=vol-521d803a

8. LAUNCH YOUR STARCLUSTER INSTANCE
(E.g., 'smallcluster')

starcluster -c /full/path/to/config start smallcluster


Hope that helps?

Cheers,

Stuart







On 2/3/2011 5:13 PM, Archie Russell wrote:
>
> Hi,
>
> Thanks for the help so far guys, I got starcluster to fire up an AWS
> cluster (config file needed a strategic " ")
>
> I am trying to mount a volume now and getting this error:
>
> clustersetup.py:200 - WARNING - Cannot find partition /dev/sdz1 on
> volume vol-521d803a
> clustersetup.py:202 - WARNING - Not mounting vol-521d803a on /bioreference
> clustersetup.py:204 - WARNING - This either means that the volume has
> not beenpartitioned or that the partition specifieddoes not exist on
> the volume
>
> I've mounted this volume before and it worked OK, but never dealt with
> partitions. What should I do?
>
> Thanks,
> Archie
>
>
> _______________________________________________
> StarCluster mailing list
> StarCluster_at_mit.edu
> http://mailman.mit.edu/mailman/listinfo/starcluster
Received on Fri Feb 04 2011 - 00:34:08 EST
This archive was generated by hypermail 2.3.0.

Search:

Sort all by:

Date

Month

Thread

Author

Subject