ORBIT-USER: Imaging Entire Grid ???

[Haris Kremo]

are 1-1 and 1-2 under working nodes just a placeholder?

we should probably organize the lists by rows or columns to make this
easily readable

H.

On 5/7/07, Ivan Seskar <Seskar at winlab.rutgers.edu> wrote:
> Hi Shivesh,
>
> Unfortunately what you are observing are deteriorating power supplies
> (or to be more specific problem with standby power);  there is not much
> we can do about it until May 20th when we are supposed to get first
> batch of replacement power supplies (manufacturer acknowledged the
> problem and agreed to ship us new ones). Now how long it will take to
> replace all of them is something we can't estimate at the moment; we
> will start with the nodes with dead power supplies (red ones one the
> status page) and slowly move to ones that are flaky (hint for the Orbit
> users who are within driving distance: any help will be greatly
> appreciated and will most certainly speed up replacement). Until then,
> best advice we can offer is to try to build the list of reliable nodes
> (i.e. ones that image properly) and restrict your experiments to that
> subset. Towards that goal, how about if everybody starts reporting nodes
> issues at:
>
> http://www.orbit-lab.org/wiki/NodeIssues
>
> Ivan.
>
> -----Original Message-----
> From: owner-orbit-user at winlab.rutgers.edu
> [mailto:owner-orbit-user at winlab.rutgers.edu] On Behalf Of Shivesh
> Makharia
> Sent: Sunday, May 06, 2007 7:49 PM
> To: orbit-user at winlab.rutgers.edu
> Subject: ORBIT-USER: Imaging Entire Grid ???
>
> Hi,
>
> I had a query regarding imaging the entire grid. Probably people who
> work on a lot of nodes might have some idea ..
>
> Whenever I try imaging the entire grid, on average only an 100 ~ 150
> nodes seem to get imaged. I also shut down all the nodes before running
> imageNodes4
>
> Does anyone have some kind of tweak (like some sequence to follow) when
> imaging the entire grid which might help in getting more nodes imaged.
> The other problem is that which so many nodes not actually being imaged,
> it is very difficult to make out which nodes have been imaged and which
> nodes have not been imaged.
>
> (I land up spending the better half of my slot finding nodes which
> actually have booted in the correct image :( )
>
> Any suggestions would be greatly appreciated.
>
> Thanks !
>
> Shivesh
>
> --
> " I arise in the morning torn between a desire to improve the world and
> a desire to enjoy the world. This makes it hard to plan the day "
> - E.B. White
>
>