[phenixbb] Lost my job history, can it be recovered? (GUI)

[Nathaniel Echols]

On Wed, Aug 1, 2012 at 7:12 PM, Andrew Waight <waight at msg.ucsf.edu> wrote:
>      It just so happened that my lab's server ran out of space (someone else
> uploading datasets) while my phenix rosetta MR job was running. Now the GUI
> no longer shows any job history in the job history tab (although on the
> title page the number of jobs is correctly displayed. I cannot find any
> information on how or if its possible to recover my job history in the GUI.
> Any tips? Thanks so much.

In theory, yes, this is possible - it really depends on what you've
done since the disk space ran out.  The GUI creates a .phenix
subdirectory in the project directory, which is where all of the
metadata, config files, and stored results are kept.  It will look
something like this:

nat at iMac:~/data/lysozyme$ ls -l .phenix/
total 96
drwxr-xr-x   2 nat staff    68 Jan 21  2012 defaults
-rw-r--r--   1 nat staff 27001 Feb 15 14:35 job_history.phil
-rw-r--r--   1 nat staff 26989 Feb 15 14:35 job_history.phil_1
-rw-r--r--   1 nat staff 26508 Feb 15 14:35 job_history.phil_2
-rw-r--r--   1 nat staff  1980 Feb 15 14:35 project.phil
-rw-r--r--   1 nat staff  1980 Feb 15 14:35 project.phil_1
-rw-r--r--   1 nat staff  1980 Feb 15 14:35 project.phil_2
drwxr-xr-x 175 nat staff  5950 Feb 15 14:35 project_data
drwxr-xr-x  37 nat staff  1258 Feb 11 14:12 tmp

The files ending in _1 or _2 are the backups.  If one of these looks
significantly larger, you want to replace the live copy with that one.
 Unfortunately, the backups are frequently replaced, so recovery needs
to be done quickly after the crisis occurs.  (I should probably add
code to check for a large discrepancy in file sizes, and automatically
sequester the backup copies in such cases.)

-Nat