Update Documentation authored by Stefan Borowski's avatar Stefan Borowski
...@@ -257,10 +257,25 @@ Ramses replicates the filesystem setup of CHEOPS, specifically: ...@@ -257,10 +257,25 @@ Ramses replicates the filesystem setup of CHEOPS, specifically:
- created on request - created on request
- NO AUTOMATIC BACKUP - NO AUTOMATIC BACKUP
### 4.1 Backing up your data ### 4.1 Archiving your data
We are working on filling in information about data backups on Ramses👷. While we back up home directories, you should archive the data in projects directories. To archive data in your project directory, you first need a TSM registration for archiving. With the registration, you are provided a nodename referring to your archive space. Then, you can access this archive space with the TSM client on the login nodes ramses[1-4].itcc.uni-koeln.de. Before archiving of data you should create a personal configuration file dsm.opt with the following content:
servernameadsm4n
virtualnodenodename
subdiryes
The server for archiving is named adsm4n. As virtual node, you should provide your personal nodename. For simple data management, subdirectories are to be included. The TSM client can be launched with the command line user interface dsmc or the graphical user interface dsmj. For archiving, you should use the command line client:
dsmc archive /projects/project/tobearchived/ \
-des=archivename -optfile=/path_to_dsm.opt/dsm.opt
The trailing slash on the directory is crucial for dsmc to recognize: it is a directory not a file. The archive name archivename will help you finding data to be retrieved from a specific archive. For retrieving data, you should use the graphical client
dsmj -optfile=/path_to_dsm.opt/dsm.opt
Within the graphical user interface, you can browse the content of your archives for data to be retrieved. For both command and graphical user interface, an absolute pathname of dsm.opt is needed (with slash in front).
NOTE: The purpose of an archive is different from that of a backup. A backup saves frequently changing data. However, an archive saves data that does not change anymore (e.g. results). Therefore, you should not archive your whole project directory but finished subdirectories. You can retrieve this data on any computer with a TSM client (e.g. workstation or laptop) for further processing.
Links:
http://rrzk.uni-koeln.de/tsm.html
https://www.ibm.com/docs/en/spectrum-protect/8.1.14?topic=clients-archive-retrieve-
your-data-unix-linux
## 5. Submitting jobs ## 5. Submitting jobs
... ...
......