TIP for ATLAS:
1) how to get a data set from Tape (Somadutta)

https://rucio-ui.cern.ch/r2d2/request
in the first step, put there the dataset name, tick “container” and press search. It should find you 1 dataset.
Select it and click on continue.
In the second step, we have to choose where we want our files be replicated. It doesn’t really matter as long as it has “SCRATCHDISK” in its name. (I guess we don’t have right to replicate anywhere else than to scratchdisks.) for us in USA, we can pick up BNL-OSG2_SCRATCHDISK, it’s large and doesn’t have problems very often.
Then click on “check quota” and you’ll see the list of sites (well, only 1 in this case) and what’s the quota. As long as we have some remaining quota, we can replicate there datasets as you want. Then just click continue again.
In step 3, there are some options, but it’s not so important. Only important thing to remember is that  datasets in a scratchdisk holds only 15 days, so we have to download it in the next 2 weeks, or there is no guarantee the dataset will be there and we may need to do this all over again. Just click continue.
In step 4, there is just an overview of what is going to happen. If it’s OK, click on “submit request”.

After that, the dataset has to be copied from tape to another site. In can take a few hours as well as few days. we can check again with “rucio list-dataset-replicas” to see where it is and whether it’s complete. If it’s complete, download it using rucio download.

2) pPb dataset from 2013 (Arabinda)