gw_data_find behaves differently than it did in the past
In the past, a gw_data_find command like this:
gw_data_find -o L -s 1238163456 -e 1240000000 -t L1_R
would return all the paths where that data lived, including duplicates like the same file living in /archive and in /hdfs (to pick a CIT specific example) and both file:// and gsiftp:// versions.
To get only one "preferred" version of each file (where the preference is specified by a data find server configuration), we'd overloaded the meaning of "-u file":
gw_data_find -o L -s 1238163456 -e 1240000000 -t L1_R -u file
But today, I see that both of these return the exact same output (the pruned list). I'm not necessarily objecting to that (we'd even talked about making it the default because it's what people mostly want and where I'd like to steer them anyway), but it does seem that there's no way to get the unfiltered list any longer.
More in the realm of "bug", the following:
gw_data_find -o L -s 1238163456 -e 1240000000 -t L1_R -m archive
used to list all paths that matched the search "archive". Now it only gives you the ones that match archive that don't have corresponding versions in /hdfs. I.e. it's now the same as you get from the old set up if you ran:
gw_data_find -o L -s 1238163456 -e 1240000000 -t L1_R -u file|grep archive
Which I don't think is the right thing to do. Undocumented feature(s)? Bug(s)? Let me know...thanks!