[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[LDM #OPH-525509]: system question

Hi Frank,

re:
> It looks like our LDM machine will be moved to a virtual cloud-based
> system here at UMass Lowell.

OK, interesting.

re:
> What kind of memory and storage would you folks recommend?  I'm hoping
> to start getting at least some of the NexRad data, as well as the new
> GOES images from Wisconsin.  Figure on a weeks worth of rotating data,
> and then have room for case study storage.

OK.

re:
> Would you recommend something like 12 gb of memory and a tb of storage?

I think that the TB of storage might be sufficient, but depending on what
else your virtual machine is doing, 12 GB of RAM might be on the low side,
20 would be likely be better.

Questions:

- how many virtual CPUs would the VM have?

  If you are planning on doing a lot of data decoding, you will need a
  beefier VM which, in turn, means more and faster vCPUs.

- what will the underlying file system be composed of?

  Meaning spinning disks or SSDs?

  I ask because we just went through an exercise troubleshooting the UNebraska's
  poor performance in a virtual machine that they were running.  It turned out
  that the underlying SSDs were basically worn out, so writes to them were 
taking
  longer and longer.  After we figured that this was the likely problem, they
  created and mounted a large (2TB) partition composed of spinning disks, and
  their performance problem went away.  The lesson learned was/is that SSDs,
  while nice and fast when new, have limited lifespans that may well never be
  reached with casual use, but can be an issue when there is a LOT of disk
  activity like is typical for machines running LDMs and decoding data.

- what kind of cloud will UMass-Lowell be running?

  I ask because quite some time ago, I interacted with a user at UChicago
  who was trying to run an LDM in the UChicago cloud, and he was encountering
  significant problems with writing the products received by the LDM to
  disk, and the bottleneck was in the underlying system as far as I could
  tell.

Cheers,

Tom
--
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: OPH-525509
Department: Support LDM
Priority: Normal
Status: Closed
===================
NOTE: All email exchanges with Unidata User Support are recorded in the Unidata 
inquiry tracking system and then made publicly available through the web.  If 
you do not want to have your interactions made available in this way, you must 
let us know in each email you send to us.