Hi,
yesterday at the meeting I said I had problems with the atlas.cern.ch
repository while upgrading 2 or 3 nodes to 2.0.3. I looked at it a bit
better and the reason only atlas was giving problems is because only
atlas jobs were running on those nodes. Although this means that the
upgrade procedure (i.e. upgrade rpms, restartautofs) works, the
conclusion is the same: it's not possible to upgrade cleanly while jobs
are running at least this time.
cheers
alessandra
|