Notes about automated archival

It's the modification time that counts. Here's an example of how it looks:

# what time is it now?
 hajaalin@dx1-biotek973:/mnt/lmu-active-rw/LMU-active1/harri/findtest$ date
 Tue Oct  7 18:08:48 EEST 2014

# create a file (test1)
 hajaalin@dx1-biotek973:/mnt/lmu-active-rw/LMU-active1/harri/findtest$ touch test1

# look at the file's timestamps: as you see birth time is not set
 hajaalin@dx1-biotek973:/mnt/lmu-active-rw/LMU-active1/harri/findtest$ stat test1
   File: ‘test1’
   Size: 0             Blocks: 0          IO Block: 16384  regular empty file
 Device: 24h/36d    Inode: 562949955276125  Links: 1
 Access: (0600/-rw-------)  Uid: (1028227/hajaalin)   Gid: (    0/    root)
 Access: 2014-10-07 18:08:58.154383300 +0300
 Modify: 2014-10-07 18:08:58.154383300 +0300
 Change: 2014-10-07 18:08:58.154383300 +0300
  Birth: -

# wait a bit and modify the file:
 hajaalin@dx1-biotek973:/mnt/lmu-active-rw/LMU-active1/harri/findtest$ echo joo >> test1

# look at timestamps again, the modification time is updated
 hajaalin@dx1-biotek973:/mnt/lmu-active-rw/LMU-active1/harri/findtest$ stat test1
   File: ‘test1’
   Size: 4             Blocks: 1          IO Block: 16384  regular file
 Device: 24h/36d    Inode: 562949955276125  Links: 1
 Access: (0600/-rw-------)  Uid: (1028227/hajaalin)   Gid: (    0/    root)
 Access: 2014-10-07 18:08:58.154383300 +0300
Modify: 2014-10-07 18:09:16.829686400 +0300
 Change: 2014-10-07 18:09:16.829686400 +0300
  Birth: -


This can be a good or a bad thing. If you are actively working on an old document, it will not disappear from the folder where you usually find it. But if you have a dataset and for some reason you modify some files, then the modified ones will stay in active, and unmodified ones will be moved to archive. And if you now do batch processing, you will miss the ones in archive. If you want to make sure your datasets are not accidentally broken like this, you can archive them manually.