Converter Issues
Very long runs
In the beginning of CERN data taking in 2006, some very long runs have been taken with 10+ binary files of 2GB each. These runs cannot be processed on the GRID at the moment for scratch space reasons (20+ GB of input plus the same for the slcio output blasts any worker node). At DESY, jobs get killed as soon as they use 25GB of scratch space in order not to compromise the worker node. The guaranteed space on any LCG site at the moment is 15GB.
Runs which are too long to fit even the binaries into 25GB are:
300133 |
15 files |
300134 |
15 files |
300135 |
15 files |
300136 |
15 files |
300141 |
13 files |
300142 |
15 files |
300143 |
15 files |
300144 |
15 files |
300145 |
15 files |
300153 |
15 files |
300154 |
15 files |
300155 |
15 files |
300156 |
15 files |
300161 |
16 files |
300163 |
13 files |
300162 |
16 files |
300447 |
18 files |
300448 |
18 files |
300453 |
18 files |
300454 |
18 files |
300505 |
18 files |
300510 |
18 files |
300511 |
14 files |
320996 |
22 files |
--
NielsMeyer - 24 Mar 2009
Topic revision: r4 - 2009-03-25
- unknown