Archives of Former Nova Users


When users leave Nova, if they have a significant quantity of files on BlueArc, their areas will be archived to tape using sam_archive_directory_image.

First, request that the permissions be setup for archiving. Ask someone with admin privileges on the Nova machines (CS liaison, scientific server support) to run the following:

chmod -R g+rwX /path/to/user/directory

Log in as novapro and execute the following. It can take a while so a screen session is suggested.

# Log into the appropriate machine

# This process takes a long time, I suggest doing 
# it in a screen session
# or screen -x if a session has already been started

# Setup the environment

# Do the archiving
mkdir -p /pnfs/nova/archives/users/<user>/<app or ana> 
sam_archive_directory_image  -v -4  -e nova -d /pnfs/nova/archives/users/<user>/<app or ana> -s <source>

# Confirm copy was successful and image is in sam
ls -ltr  /pnfs/nova/archives/users/<user>/<app or ana>/
samweb list-files "archive.source \"<source>\"" 
samweb get-metadata <archive image>

# If it checks out, remove the source directory and add info for retrieval
rm -rf <source>/*
echo 'samweb list-files "archive.source \"<source>\""' > <source>/ARCHIVE

How to access archived files:

This is a specific example, but should be instructive.

First, locate the archive file.

$ samweb list-files "archive.source \"/nova/app/users/betan009\"" 
$ samweb locate-file image-d6577cd9-ba81-4849-96f7-3377dd22fe46.tgz

Next, check if the archive is cached or only on tape:

$ image-d6577cd9-ba81-4849-96f7-3377dd22fe46.tgz
Cached: 0    Tape: 1

If the file is already cached, you can skip the next step.

To cache the file from tape, do the following:

$ samweb create-definition temp-prestage-betan009 "archive.source \"/nova/app/users/betan009\"" 
$ samweb run-project --defname=temp-prestage-betan009
$ samweb delete-definition temp-prestage-betan009

When the command completes (can take some time) you can continue.

It will be fairly slow to copy over and untar the whole thing, but there is a trick to access specific files. Do as follows:

$ tar -tzf /pnfs/nova/archives/users/betan009/app/image-d6577cd9-ba81-4849-96f7-3377dd22fe46.tgz | grep "<the file you want>" 
$ tar -xzf pnfs/nova/archives/users/betan009/app/image-d6577cd9-ba81-4849-96f7-3377dd22fe46.tgz /path/to/the/file/you/want

Areas that have been archived:

  • /nova/Ana/users/cerretan to /pnfs/nova/archives/2015-May/cerretan
  • 182G /nova/app/users/denis to /pnfs/nova/archives/users/denis/app
  • 178G /nova/app/users/timdkut to /pnfs/nova/archives/uers/timdkut/app
  • 176G /nova/app/users/betan009 to /pnfs/nova/archives/uers/beta009/app
  • 200G /nova/app/users/blinehan to /pnfs/nova/archives/users/blinehan/app

Areas that need to be archived:

  • 128G /nova/app/users/rocco
  • 113G /nova/app/users/xbbu
  • 108G /nova/app/users/waldron
  • 113G /nova/app/users/kkelly12
  • /nova/app/users/tamsett
  • /nova/app/users/raddatz
  • /nova/app/users/dmitryr
  • /nova/app/users/rtoner
  • /nova/app/users/avgavril
  • /nova/app/users/mercurio
  • /nova/app/users/kravtsov