60 likes | 179 Vues
This overview details the current open issues identified during tests of the WP1 WMS Release 2. While several minor UI problems and resubmission issues have been noted, the team is actively addressing them. Outstanding concerns relate to Condor-G job submissions, such as idle states upon Globus submission failures and proxy expiration notifications. Furthermore, integration tasks with WP2 components are ongoing. Essential documentation updates and testing of new functionalities are still pending, ensuring all elements function seamlessly. Feedback on additional issues is welcome.
E N D
WP1 WMS release 2 issues Massimo Sgaravatto INFN Padova
Disclaimer • The issues that I am going to raise come from tests I performed • I didn’t tests yet all the functionalities and all the components • Please let us know if there are other open issues
Open issues • Some “last minute” UI problems • Trivial to fix • Problems with resubmission • CEs already “used” are not considered • Some open issues with Condor-G • In some cases when submission to Globus fails, job stays idle, and nothing is reported • Nothing is reported when the proxy expires • … going to be fixed soon • Some problems affecting UI with JDLs logged to LB • Debugging messages to be removed in UI/NS client
Missing pieces • Dynamic quota management in NS • Maradona • Ready in LM, still missing in JA • Integration with WP2 Replica Manager • Integration with RLS (ListReplicas) • ~ done, but not tested yet (since RLS not populated with significant [for us] entries) • Integration with Optor (getaccesscost as rank) • To be done in release 2.0 or later ? • Gangmatching ? • Status ?? • LB • Interlogger crashes(d) very often • Hopefully all problems fixed
Missing pieces • Output Data Registration • To be done in release 2.0 or later ? • Management of expired proxies • We have to decide what to do • Also need to understand how Condor-G is going to manage these scenarios • To be done after release 2.0, I suppose • UI man pages • Documentation • WMS user and administrator guide • Admin guide ~ ready (apart from UI part) • User guide update on going • Build part missing • Need to explain somewhere the new extended LB querying capabilities
Testing • These issues come from tests I performed only using the python UI • Tested “old” functionalities • New functionalities • Quickly tested MPI jobs both on LSF and PBS • Patched gridftp server and gridftp API tested • I didn’t test yet GUI and APIs • I didn’t test yet the other new functionalities • Proxy renewal, Interactive jobs, Quota management for sandbox files, Job Checkpointing, purger daemon, … • No stress tests