-- DavidHammer - 06 Jan 2010

This list was generated simply by looking at the newest and oldest canonical_results directory. The other results (valid, weird, ...) should be in the same location.

TODO From Bruce (add example filenames to list)

In addition to the dates and directory names, it would be worth also listing the names of a few of the typical results. From the file names we can tell which 'run' of E@H searching it was. This is especially useful for the oldest results on einstein-fs1.phys.uwm.edu, which are from before the rewritten assimilator that put things into directories named by the search run.

einstein.phys.phys.uwm.edu
  • /raida/storage/S5R6/results/
    • canonical_results_2010_01_06
    • canonical_results_2009_10_14
    • 106 days directories
  • /raida/storage/S5R5/results/
    • canonical_results_2010_01_02
    • canonical_results_2009_08_18
    • 105 days directories
einstein-fs3.phys.uwm.edu
  • /export1/storage/S5R5/results/
    • canonical_results_2009_08_11
    • canonical_results_2009_01_14
    • 205 days directories
  • /export1/storage/S5R4/results/
    • canonical_results_2009_05_24
    • canonical_results_2008_07_28
    • 253 days directories
  • /export1/storage/RESULTS_DAILY
    • canonical_results_2008_09_25
    • canonical_results_2007_06_13
    • 471 days directories
  • /export1/storage/ABP1/results
    • canonical_results_2009_03_27
    • canonical_results_2008_12_18
    • 33 days directories (only ~8MB total)

einstein-fs1.phys.uwm.edu
  • This fileserver contains most of the results older than the fs3 results. Einstein-fs1 has a problem with the raid array that needs to be fixed before the machine can be booted. The results will be listed here once this machine can be searched.
  • efs1 is finally running.
  • canonical_results_2005_02_05
  • canonical_results_2007_06_13
  • 856 days directories
  • the /export1 directory con 6.7T but probably only a fraction of that is data that needs to be moved to Hannover

DATA VOLUME

This is just a guess at the data volume. There are 669 daily directories for S5R6 and S5R5 and each canonical_results dir is about 8GB. This is 5TB for just canonical results of S5R6 and S5R5.

BACKUP PLAN

Bernd sent an email to set the priorities.

  1. canonical results
  2. transcripts of purged WUs and results (in /BOINC/projects/EinsteinAtHome/archive) (!)
  3. valid results (to allow error correction in case something goes wrong with > one or the other canonical result)

Plan

  • New canonical results will be backed up by $E/bin/rsync_hannover/rsync.sh run under boinc cron (config.xml)
  • Older canonical results will be backed up by a script that uses grid tools and will be run by hand.

STATUS

2010-02-22

einstein.phys.uwm.edu
  • S5R6 (106+ days)
    • canonical DONE
    • valid DONE
  • S5R5 (105 days)
    • canonical DONE
    • valid DONE

einstein-fs3.phys.uwm.edu
  • S5R5 (205 days)
    • canonical DONE
    • valid DONE
  • S5R4 (253 days)
    • canonical DONE
    • valid DONE
  • RESULTS_DAILY (471 days)
    • canonical DONE
    • valid DONE
  • ABP1 (33 days only 8MB total)
    • canonical DONE
    • valid DONE

  • There are 2280 directories to move.
  • So far 1538 have been moved (67%).
  • 742 directories to move yet
  • 436 directories were moved in the last two weeks [ find RESULTS_DAILY S5R* -mtime -14 |grep tar | wc -l ]
  • moving directories at rate of 31 / day
  • expected to be done on March 18 (in 24 days)
Topic revision: r14 - 20 Apr 2010, DavidHammer
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback