[StarCluster] StarCluster not mounting EBS volume correctly

[Dan Tenenbaum]

One more thing. After successfully launching a cluster and having it connect
to the ebs volume, I shut the cluster down.

Then I started it again, this time I got:

>>> Validating cluster template settings...
cluster.py:766 - ERROR - volume vol-70e6f719 is not available (status:
in-use)
cli.py:243 - ERROR - settings for cluster template "smallcluster" are not
valid

I went to the AWS console and detached the volume (funny how a volume can be
attached to an instance that has terminated). Then I was able to start my
cluster again without error.

Is this a bug? Shouldn't the volume be detached as part of "starcluster
stop"?

Dan


On Wed, Oct 27, 2010 at 1:22 PM, Dan Tenenbaum <dtenenba at fhcrc.org> wrote:

> Hi Justin,
> I formatted the volume according to your instructions below and everything
> worked.
>
> BTW, everything done with ElasticFox in those instructions can now be done
> in the AWS console (which is how I did it). For those who don't use firefox.
> ;)
>
> Dan
>
>
> On Wed, Oct 27, 2010 at 12:44 PM, Justin Riley <jtriley at mit.edu> wrote:
>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Hi Dan,
>>
>> The 0.91 version of StarCluster assumes the volumes are partitioned with
>> a single partition. I know now that it's best *not* to partition the
>> drives but rather just format the raw disk. Unfortunately for now using
>> partitioned EBS vols will have to do.
>>
>> I'm currently working on updating the code to support both partitioned
>> and unpartitioned vols along with a slew of other fixes and enhancements
>> but in the mean time I would recommend using StarCluster's createvolume
>> command to create, format, and partition a new EBS volume for use with
>> StarCluster:
>>
>> http://web.mit.edu/stardev/cluster/docs/volumes.html
>>
>> HTH,
>>
>> ~Justin
>>
>> On 10/27/2010 03:29 PM, Dan Tenenbaum wrote:
>> > Hi all,
>> >
>> > I created a 30GB EBS volume, formatted it with the ext3 filesystem, and
>> > detached it.
>> >
>> > Then I set my cluster config to use this volume.
>> >
>> > VOLUMES = sharedhome
>> > ...
>> > [volume sharedhome]
>> > VOLUME_ID = vol-70e6f719
>> > MOUNT_PATH = /home
>> >
>> > I start my cluster and see the following output:
>> >
>> >>>> Waiting for cluster to start.../>>> The master node is
>> > ec2-184-73-92-19.compute-1.amazonaws.com
>> > <http://ec2-184-73-92-19.compute-1.amazonaws.com>
>> >>>> Attaching volume vol-70e6f719 to master node on /dev/sdz ...
>> >
>> >>>> Setting up the cluster...
>> >>>> Mounting EBS volume vol-70e6f719 on /home...
>> > ssh.py:77 - WARNING - specified key does not end in either rsa or dsa,
>> > trying both
>> >>>> Using private key /Users/dtenenba/.ec2/bioc-keypair.pem (rsa)
>> > clustersetup.py:201 - WARNING - Cannot find partition /dev/sdz1 on
>> > volume vol-70e6f719
>> > clustersetup.py:203 - WARNING - Not mounting vol-70e6f719 on /home
>> > clustersetup.py:206 - WARNING - This either means that the volume has
>> > not beenpartitioned or that the partition specifieddoes not exist on the
>> > volume
>> >
>> > I formatted my volume according to these instructions:
>> > https://learningspaces.njit.edu/content/how-format-and-mount-ebs-volume
>> > Was there something else I need to do? How do I put a named partition
>> > such as /dev/sdz1 on the volume?
>> > BTW, I don't want to partition the volume, but I want to use the whole
>> > thing.
>> >
>> > I have set up volumes before using this procedure and not run into any
>> > problems.
>> >
>> > BTW, after I ssh into the master, and do fdisk /dev/sdz, I see this:
>> > # fdisk /dev/sdz
>> > Device contains neither a valid DOS partition table, nor Sun, SGI or OSF
>> > disklabel
>> > Building a new DOS disklabel with disk identifier 0x5cde0278.
>> > Changes will remain in memory only, until you decide to write them.
>> > After that, of course, the previous content won't be recoverable.
>> >
>> >
>> > The number of cylinders for this disk is set to 3916.
>> > There is nothing wrong with that, but this is larger than 1024,
>> > and could in certain setups cause problems with:
>> > 1) software that runs at boot time (e.g., old versions of LILO)
>> > 2) booting and partitioning software from other OSs
>> >    (e.g., DOS FDISK, OS/2 FDISK)
>> > Warning: invalid flag 0x0000 of partition table 4 will be corrected by
>> > w(rite)
>> >
>> > Not sure if that is relevant or not.
>> >
>> > Please let me know how I am supposed to configure the volume. Thanks
>> >
>> > Dan
>> >
>>
>> -----BEGIN PGP SIGNATURE-----
>> Version: GnuPG v2.0.16 (GNU/Linux)
>> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
>>
>> iEYEARECAAYFAkzIgTsACgkQ4llAkMfDcrmelwCgk7a/KNmr07niWF8Ibh0CZZlR
>> 8qAAn2HfyqvfoAZjYzmn/aIbPAmzElXt
>> =/o9a
>> -----END PGP SIGNATURE-----
>> _______________________________________________
>> StarCluster mailing list
>> StarCluster at mit.edu
>> http://mailman.mit.edu/mailman/listinfo/starcluster
>>
>>
>
> _______________________________________________
> StarCluster mailing list
> StarCluster at mit.edu
> http://mailman.mit.edu/mailman/listinfo/starcluster
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/starcluster/attachments/20101027/bef21508/attachment.htm