Collaboration meeting
Vidyo
# Argus Collaboration Meeting - May 20, 2016
[Agenda](https://indico.cern.ch/event/522011/)
## General News
NTR
## New Release
Last issue with mapping of secondary FQANs fixed (primary FQAN considered as a secondary FQAN)
* Fix should be available today
* As a regression not covered by tests: fixed also
Release contents ready
* Including startup script fixes
* Release numbering fixed: ready for release
Packages available for EL6 and EL7
* Tested successfully
Still missing: release notes
CERN ready to deploy in production a few instances with the new release and assess everything is correct
* After Q/A assessment the plan is to move all the instances to the new version to avoid the hacks for mapping
consistency between new and old release
* The CERN mapping problem should not affect other sites: specific to CERN use of gridmapfile
* New code expecting to put a much lower load on the (NFS) backend: would be good to assess if NFS attribute caching
(enabled as a workaround with the old code) could be disabled without visible effect on the service.
Maarten suggests that some source files are no longer used and that some cleanup may be needed
* To be followed-up... Maarten will try to provide evidence
Stress testing by NIKHEF: problem running the tests (private key required but not handled)
PIP for IOTA CA: code expected to be ready by the end of the summer for integration into the PEPd code
* In fact split into 2 PIPs
* Plan to run performance tests
### UMD4 Release
Need to sort out how to do the staged rollout
* But CERN deployment should mitigate the importance
Need also to fix the Puppet module
* In particular, the replacement of the emi metapackage: should be easy to fix!
* Andrea: there is also a new metapackage available
* Everybody surprised by this becoming a prerequisite for the release
* Previously agreed that the requirement was clear installation instructions that are available
* In the case of EL6, this is just an upgrade
Vincenzo agrees to move forward and wait for a later release the possibility of a standalone Puppet à la DPM.
## Documentation
Waiting for last cleanups for merging the PR. Should be ready soon.
* Agreement to remove the information about very old version
## Other Issue
Mischa: no reference in the log about the MD5 checksum no longer available
* A long GGUS ticket that turned out to be problem caused by a very old CREAM client used in HTCondor still using MD5 checksums: troubleshooting turned out to be very cumbersome
* Andrea: difficult to get the information, in fact non trivial...
* Can add this issue into the troubleshooting of the documentation
In general, we should try to follow more closely the discussion about deprecated features related to this
security topics to better anticipate problems.
## AOB
Next meeting: June 24, 2pm