Data on old CSAIL NFS servers must be moved

The following information is provided for historical documentation. The old NFS servers have now been shut down, and references to them have been removed from the automount maps.

All data on the old CSAIL NFS servers must be reorganized and relocated to new hardware by the end of the Spring '13 term. If your data is not completely moved by then, you will be presumed not to care about it any more, and it will be taken offline. These servers, and the disks attached to them, are approaching end-of-life and no longer have support from the manufacturer.

Which data is affected?

This change affects only NFS servers (usually mounted under /data or /csail on CSAIL machines), and of the NFS servers, only the three old ones: zim, gir, and phobos. This includes the following (sizes as of 2013-04-02):

Group/Project zim gir phobos
billf 3515 GB 5324 GB 8396 GB
cag - 2662 GB 1946 GB
cocosci - - 497 GB
dgc - 5734 GB -
dogs 0.02 GB - -
fisher 3982 GB 1126 GB 1126 GB
infolab - - 2970 GB
lsp 2458 GB - -
marine 1126 GB - -
nlp 4936 GB 969 GB -
oliva - - 1946 GB
polina 14336 GB 3686 GB 8397 GB
rrg 2253 GB - -
sls - - 3174 GB
tidor - - 254 GB
torralba 10854 GB 14336 GB 9318 GB
videolabelme - - 13312 GB
vl 98 GB - -
wearable - - 83 GB
wiglan 132 GB - -
wikido 640 GB 2867 GB 568 GB

What needs to be done

We need to break your existing data storage down into manageable, homogeneous chunks. We can do this by project, or by type of data, or by dataset (if your datasets are reasonably large), but we need you to tell us:
  • What are those chunks (name, approximate size, and expected usage)?
  • Where should we copy the data from? (Or if you prefer, you can do the copy yourself.)
  • Is it compressed or encrypted?
  • What backup policy is required? (See NFS Snapshots and Backups for more details.)
  • If this is archival data, can it be packed up into an archive format?

For additional background on requesting NFS filesystems, see NFS Storage.

If you need a scratch space above and beyond the public /data/scratch, let us known and we can probably provide that as well, on a separate server.

-- GarrettWollman - 02 Apr 2013
Topic revision: 01 Jul 2013, GarrettWollman
 

MIT Computer Science and Artificial Intelligence Laboratory

 

  • About CSAIL
  • Research
  • News + Events
  • Resources
  • People

This site is powered by Foswiki MIT: Massachusetts Institute of Technology