Ciao Alessandro,
>> In particular, sites supporting ATLAS or CMS are advised _not_ to deploy
>> the EMI-1 Worker Node until further news later this month, during which
>> we will try to assess how to deal with issues that have affected or may
>> affect some of the experiment software in that new environment.
> Hi Maarten,
> do you have any pointers to these problems?
> Just to inform the sites (and also we NGI managers...) about the problem
> that is affecting the software of some VOs
There are at least the following:
https://ggus.eu/ws/ticket_info.php?ticket=78315
(CMS software has issues running on EMI WN)
https://ggus.eu/ws/ticket_info.php?ticket=79727
(lcg-cr failure in emi-wn-1.0.0-0.sl5.x86_64)
https://ggus.eu/ws/ticket_info.php?ticket=79093
(lcg_cr gives Internal error)
https://ggus.eu/ws/ticket_info.php?ticket=78797
(segmentation fault when using lcg-cr)
In the coming weeks we will try to assess to what extent the affected
experiments can work around such matters.
> in the past months we pressed a lot the sites to migrate to EMI, and now
> this news leave us quite dismayed :-)
The SAM tests only show a small number of small sites reporting an EMI
version on the WN. Some may have set up an EMI WN just for "ops"...
|