search

fridoling / sequencing

0 stars 0 forks source link

writing permission to the /sequencing folder on the cluster #4

Open pilopilota opened 5 years ago

pilopilota commented 5 years ago

I can't rsync the git folder to the cluster, since I don't have permission to write in the folder. maybe you can change this. thanks!

pilopilota commented 5 years ago

also the exp_small folder into DataTest has the same issue. I'm not planning to write in it now, but in case...

fridoling commented 5 years ago

Sorry about that. I changed the permissions to 775 everywhere. But you wanted to rsync the git folder from the cluster to somewhere else, right? On the cluster it should only be updated using git pull.

pilopilota commented 5 years ago

aah, I thought we'd use rsync on the cluster. I'm fine with git pull. sorry

fridoling commented 5 years ago

I think it's necessary because some parts of the script only run on the cluster and I cannot test them on my computer (java issues!) so I want to be able to make changes to the scripts directly on the cluster and then synchronize them to the github repository. BUT: I am not sure now whether we can both work on the same repository on the cluster. Maybe it would be better if each of us has their own separate clone of the repository on the cluster for testing and developing (and running?) the scripts. Do we in addition need a third 'impersonal' version of the repository on the cluster? Maybe not...

pilopilota commented 5 years ago

mmm at the moment I'm not working on the cluster but only on my local files and then git push. I can test my scripts on my computer, I don't need a personal folder on the cluster. if you prefer to have a personal folder on the cluster for testing, I'm fine with it, otherwise for me it is ok as we're doing now: git pull on the common cluster folder.

fridoling commented 5 years ago

Have you already tried git pull on the cluster? Maybe it works since you are a collaborator. But we have to make sure that we do not access it at the same time. So it's maybe best if I create another clone of the repository in my home on the cluster and the 'impersonal' clone in /lustre/data/ANC/NGS will only be used for pulling and not for pushing.

fridoling commented 5 years ago

Sorry for spamming, but maybe it would be best if you also have a personalized version of the repository on the cluster just to run the scripts. Otherwise git will recognize any change of parameters (e.g. the foldernames) and those changes will have to be undone before pulling a newer version. I don't know what the best way is. Maybe we can decide tomorrow.