 |
 |
Hi Jeff, Like I said in the previous mail. This assumption had to be made due to a problem with the Glue Schema ( and maybe a limitation in some batch systems). People are wanting to do accounting on a per VO basis and this was the short term solution. It has absolutely nothing to do with CERN at all. CERN is facing the same problem as all the other sites as they only have one grid queue for the LSF batch system. The correct solution is to fix the real problem and this is what we are trying to do. Steven Burke is currently finishing of a document that suggests this fix. As soon this is been done we will try and get the schema modified. Laurence Jeff Templon wrote: > Yo, > > On Tue, 2005-01-18 at 14:25, Laurence wrote: > > > The reason why this assumption has been made is due to a problem in the > > Glue Schema that we are attempting to fix. This is the only way > that we > > can get accounting to work correctly. We need to know the number of > jobs > > running and waiting per VO. > > Bad bad bad. I suggest this be brought up at the next LCG workshop, and > possibly at the GDB as well. > > I have sympathy for the problem, but we should not fix problems in the > middleware by making the lives of 100+ people (the sysadmins) > difficult. And especially not without agreeing with these people > beforehand. CERN has a totally different set of constraints than most > sites. > > Absolutely unacceptable: "configure your site like CERN does, or else > the software will not work". Next we will be told "install CERN > Scientific Linux or else the software will not work". What's next, > "send us your computers and we will install them and run them here??" > > This is why the authors of EGEE JRA1 part of the proposal suggested to > NOT have the integration team at CERN. That is back when we thought > that JRA1 would be doing this, not LCG ;-) > JT >