[Starcluster] problem with centOS ami?

[Dan Yamins]

On Sun, Sep 5, 2010 at 10:10 PM, Fred Rotbart <rotbart at softbart.com> wrote:

> Hi Dan,
>
> The image I meant was ami-7ea24a17 which is the Amazon generic Cluster
> Compute image.
> Rightscale have a beta image,ami-78947f11, which is also Cluster Compute
> compatible.
>
>
Ah -- well, I wasn't intending it to be a Cluster Compute image ... it was
just meant for "regular" amazon images.   Are you asking about building a
centOS Cluster Compute image?



> In fact, AFAIK currently these are the only two public images available for
> Cluster Compute.
>
> Any thoughts?
>
> - Fred
>
>
> On 3/09/10 23:57 , Dan Yamins wrote:
>
>>
>>
>> On Thu, Sep 2, 2010 at 9:53 PM, Fred Rotbart <rotbart at softbart.com
>> <mailto:rotbart at softbart.com>> wrote:
>>
>>    Hi Dan,
>>
>>    Thanks for the help.
>>
>>    I had started building a lot of stuff from sources. However, the
>>    Python 2.6 I took from
>>    http://chrislea.com/2009/09/09/easy-python-2-6-django-on-centos-5/.
>>    Any advantage to building the Python from sources?
>>
>>
>> Well, three reasons:
>>
>> 1) I needed a 64-bit install.   Typically I've found that 64-bit
>> pre-compiled binaries are less reliable than building 64-bit.   This is
>> the main reason I used a source as opposed to a pre-compiled binary (e.g
>> from python.org <http://python.org> itself).
>>
>>
>> 2) As explained at the top of Chris Lea's post, overwriting the system
>> python on centOS with python26 is really bad.   I really really needed
>> to make sure that it didn't screw things up and so I decided to install
>> in /opt, where a completely standard install would work fine and be sure
>> not to interfere.  To make a package install work while completely not
>> interfering with the other system installs, putting all the relevant
>> byte-compiled dependencies in version-number-named places throughout the
>> distros, requires alot of finicky work. I'm not comfortable enough with
>> yum to feel like I totally understood how to evaluate whether any given
>> person's repo was likely to be "right".    This is the main reason I
>> would use a non-yum-based install vs. a repo install.
>>
>> 3) I wasn't aware of Chris Lea's work -- thank you so much for telling
>> me of it  If I had known of it, perhaps I would have used it (if there's
>> a 64-bit version) -- but subject to the above provisos.
>>
>>
>>    BTW why did you use the Rightscale image rather than using the
>>    Amazon one?
>>
>>
>> I've typically found Rightscale images to be more reliable than Amazon
>> ones actually -- and more things I already need are on them.   Which
>> amazon image  were you referring to?
>>
>>
>>
>>    - Fred
>>
>>
>>    On 3/09/10 09:49 , Dan Yamins wrote:
>>
>>        Fred, sorry for the delay in reply.
>>
>>        I've put the instructions up here at this google doc
>>
>>
>> https://docs.google.com/Doc?docid=0Aa_xSFd7IOCfZGhzN3JwanBfMTA1ZjRwNWp6dGo&hl=en&authkey=CI2p-t8J
>>        <
>> https://docs.google.com/Doc?docid=0Aa_xSFd7IOCfZGhzN3JwanBfMTA1ZjRwNWp6dGo&hl=en&authkey=CI2p-t8J
>> >
>>        <
>> https://docs.google.com/Doc?docid=0Aa_xSFd7IOCfZGhzN3JwanBfMTA1ZjRwNWp6dGo&hl=en&authkey=CI2p-t8J
>>        <
>> https://docs.google.com/Doc?docid=0Aa_xSFd7IOCfZGhzN3JwanBfMTA1ZjRwNWp6dGo&hl=en&authkey=CI2p-t8J
>> >>
>>
>>        Let me know if you can't access this doc.
>>
>>        On Tue, Aug 31, 2010 at 8:35 PM, Fred Rotbart
>>        <rotbart at softbart.com <mailto:rotbart at softbart.com>
>>        <mailto:rotbart at softbart.com <mailto:rotbart at softbart.com>>>
>> wrote:
>>
>>              Hi Dan,
>>
>>            Did you happen to make a cookbook for making the CentOS
>>        StarCluster AMI?
>>            If so can you make it available?
>>
>>            - Fred
>>
>>
>>            _______________________________________________
>>            Starcluster mailing list
>>        Starcluster at mit.edu <mailto:Starcluster at mit.edu>
>>        <mailto:Starcluster at mit.edu <mailto:Starcluster at mit.edu>>
>>
>>
>>        http://mailman.mit.edu/mailman/listinfo/starcluster
>>
>>
>>
>>
>>        _______________________________________________
>>        Starcluster mailing list
>>        Starcluster at mit.edu <mailto:Starcluster at mit.edu>
>>        http://mailman.mit.edu/mailman/listinfo/starcluster
>>
>>
>>    --
>>
>>    ===================================================================
>>    Dr. Frederick C. Rotbart    Rotbart Software Technologies Pty. Ltd.
>>    Tel: +61 9024 3914          38 Arthur Street,
>>    Mobile: +61 422 569 102     Vic. 3141 Australia
>>    rotbart at softbart.com <mailto:rotbart at softbart.com>
>>
>>
>>
>>
>>
>> _______________________________________________
>> 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/20100906/cd43fd7c/attachment.htm