Present - Catherine Zwahlen (CZ), Lukasz Dutka (LD), Christian Bernardt
(CB), Jeffrey Muller (JM), Matthew Viljoen (MV)
Absent/Apologies - Huanxiang Lu (HL), Michaël Dupont (MD)
SERVICE STATUS & DATA HOSTING
- LD reported that all new data has been copied to the onedata instance
but omitted 'HBP' in the directory structure (see meeting minutes 1
Mar). This would be corrected today. Nothing further to report - service
is normal.
- CB reported that all new data has been copied to the dcache instance.
No further instability has been experienced since the latest tweaking of
dCache
- MV has reinstalled the test ATLASViewers (http://155.210.198.109/
onedata and http://155.210.198.110/ dcache) with the newest
imagestack_info.json config file created by HL. There still seems to be
a problem with these VMs being reliable - possibly due to the chosen
fedcloud site. CB asked whether all datastacks are visible to the
EGI-hosted ATLASViewer - the Golgi Rat datastack doesn't seem to be
visible. CZ said she would check whether the config file was up-to-date
-> MV to check which fedcloud site is suitable for long-term hosting of
the test ATLASViewers
-> CZ to check whether imagestack_info.json created by HL and sent to MV
by MD on 9 Mar was the latest.
- MV asked whether any further open datasets will be made available to
EGI. CZ said no - the one mentioned by LD has not been made public.
- CB asked about whether there is any documentation on the data files -
e.g. boundaries of the tile parameters. CZ said there is metadata
associated with the files and she can make point the EGI team to its
documentation.
AUTHENTICATION
- Regarding authentication, CZ confirmed that the statement from the
3 Mar meeting minutes was correct. At present, EGI will only be used to
host open datasets that do not require authentication in front of the
image services. From Apr, LD and CB can look at integrating their data
hosting services with OpenIDConnect, that will enable EGI to host closed
datasets.
- JM confimed that OpenIDConnect service was in production and
documentation on it can be provided if required. JM also said that
accounts can be setup for the EGI team. LD and MV had done this, CB will
request an account from JM.
MOVE TO PRODUCTION
- CZ requested that HBP would like EGI to host a production quality data
hosting service from 1 Apr. MV said the EGI team would need to discuss
the implications of this.
-> MV to meet with LD and CB to discuss whether we can provide a
production level service from 1 Apr - and what additional steps would be
needed to do this (e.g. monitoring, hardware resiliency etc.)
AOB
- MV will be presenting the HBP Pilot work at the Neuroinformatics session
at the EGI conference and would like to do a demo. CZ suggested EGI may
like to try a newer version of the ATLASViewer for the demo.
- MV is compiling information for the report as requested by Sean. Will
be ready by the end of the month.
DONM: 29 Mar at 10:30 CET. LD to chair (MV on holiday)