Closed dominikl closed 9 months ago
Something's a little bit strange. I ran the stats for idr0143 with the --disable-fsusage
flag. But still this number of 29Tb showed up for disk space usage, but the number of files were missing. And when I run du
on the filesystem I get 19Tb. That's quite a difference. Any idea @sbesson ?
Trying again:
(venv3) [dlindner@prod119-omeroreadwrite metadata]$ python idr-utils/scripts/stats.py --release prod119 --disable-fsusage idr0143-herbst-coculture
Using session for demo@localhost:4064. Idle timeout: 10 min. Current group: Public
idr0143-herbst-coculture screenA prod119 3452 225 86400 112 259200 3208320 29.937475584 29937475584000 2160 x 2160 x 5 x 2 x 1 cell
And with du
:
MIP plates
(venv3) [dlindner@prod119-omeroreadwrite 20221222-ftp]$ du --max-depth=1
...
2687536940
Raw plates
(venv3) [dlindner@prod119-omeroreadwrite 20220822-ftp]$ du --max-depth=1
...
15864049064
Pretty sure if you disable --fs-usage
, the script attempt to guess the volume of data based on the pixel dimensions & type i.e. the number of bytes.
This obviously ignore several considerations including the presence of pyramidal levels, the underlying file format & the usage of compression.
So it would be better to use the size from OS du
right?
So it would be better to use the size from OS du right?
Yes, there are different interpretations of what "size" is expected to capture but I think the file size captured by du
is one that is easy to interpret. For people interested in downloading the raw data, this also an informative metrics.
See title. Also added the
cell
studies type for idr0139 and idr0149, for some reason that was missed last time.