multiple users/accounts?

[cookie33]

Hi,

We store data of multiple institutes in one iRODS configuration for eudat.

Is it possible to generate several uploads of data to eudat for different insititutes and differentiate:

• institute A has x bytes and y files
• institute B has z bytes and q files

[cookie33]

The ini file is usable for only one institute

rods$ cat irodscollector.ini 
#
# template of a configuration file for EUDAT's irodscollector
#

# section containing the logging options
[Logging]
log_file=eudatacct.log

# section containing the properties to access the accounting server
# to get statistical data and report them
[Report]
# base URL of the accounting server to be used
base_url=https://accounting.eudat.eu
# domain: either eudat or test or demo
domain=eudat
# uid of the corresponding registered storage resource on DPMT 
# (same as storage_space_uuid on RCT)
account=<insert uid here>
# username of the provider on the accouniting server
# owning the account specified above
# contact dp-admin@mpcdf.mpg.de if you need one
user=<username of provider>
# if you have an access token from RCT already reuse that here
password=<password or access token>
service_uuid=<unsuported at the moment>

# section contains the list of collections to be accounted together, replace
# the examples with your collections, the script sums the values of all
# collections and sends it to EUDAT's accounting service.
[Collections]
clist=
  /zone/some/path
  /zone/other/path

We service different federations and institutes. Do I need to create different ini files for each? And start an separate collector for each federation/institute?

We have for instance:

$ ils /vzx/eudat
/vzx/eudat:
  C- /vzx/eudat/A
  C- /vxx/eudat/B
...
  C- /vzx/eudat/F

Each is a different institute within EUDAT and is registered differently within EUDAT.

I only want to run the collector once and let him do the work. And not have 5 ini files and start the collector five times.

[cookie33]

If the config is as follows it suits us more:

#
# template of a configuration file for EUDAT's irodscollector
#

# section containing the logging options
[Logging]
log_file=eudatacct.log

# section containing the properties to access the accounting server
# to get statistical data and report them
[Report]
# base URL of the accounting server to be used
base_url=https://accounting.eudat.eu
# domain: either eudat or test or demo
domain=eudat

# institutes which this iRODS instance handles
institutes=inst_a,inst_b

[inst_a]
# uid of the corresponding registered storage resource on DPMT 
# (same as storage_space_uuid on RCT)
account=<insert uid here>
# username of the provider on the accouniting server
# owning the account specified above
# contact dp-admin@mpcdf.mpg.de if you need one
user=<username of provider>
# if you have an access token from RCT already reuse that here
password=<password or access token>
service_uuid=<unsuported at the moment>

# section contains the list of collections to be accounted together, replace
# the examples with your collections, the script sums the values of all
# collections and sends it to EUDAT's accounting service.
clist=
  /vzx/eudat/A
 /vzx/other/pathA

[inst_b]
# uid of the corresponding registered storage resource on DPMT 
# (same as storage_space_uuid on RCT)
account=<insert uid here>
# username of the provider on the accouniting server
# owning the account specified above
# contact dp-admin@mpcdf.mpg.de if you need one
user=<username of provider>
# if you have an access token from RCT already reuse that here
password=<password or access token>
service_uuid=<unsuported at the moment>

# section contains the list of collections to be accounted together, replace
# the examples with your collections, the script sums the values of all
# collections and sends it to EUDAT's accounting service.
clist=
  /vzx/eudat/B
  /vzx/other/pathB

If the institutes parameter is NOT there it should behave as before. Otherwise it might have 2 configurations to check/loop about.

[raphael-ritz]

Hello again, thanks for this suggestion. In fact at MPCDF we have a similar situation. The way we are dealing with it is as you describe above (but don't like): We make as many copies of the config templates as we have resources to account on that server adapt them and then call the iRODScollector for each config file individually.

I understand that this is how it used to be before. I only ported the code from Claudio, Elena and Beatriz to work with the new accounting service.

But I like your suggestion. Should I have time one day I'll give it a spin (famous last words...)

Cheers,

Raphael

[cookie33]

A colleague made fix and we tested it:

$ iRODScollector -c /opt/eudat/b2safe/conf/irodscollector_multi.ini 
iRODScollector called with: Namespace(comment='', configpath='/opt/eudat/b2safe/conf/irodscollector_multi.ini', key='', measure_time='', object_type='registered object', service='(default)', test=False, type='storage', unit='byte', verbose=False)
Configuration file: /opt/eudat/b2safe/conf/irodscollector_multi.ini 

***clarin
Collections to be accounted:
/vzSARA1/eudat/clarin
/vzSARA1/home/eudat#vzRZGE
/vzSARA1/home/rods#replixRZG
***knmi
Collections to be accounted:
/vzSARA1/eudat/knmi
/vzSARA1/home/rods#ORFEUS
/vzSARA1/home/orfeus#ORFEUS

It will put in a pull request.