Anastasiia Chub,
Tigran Mkrtchyan
06/05/2026, 10:00
Artur Gottmann,
Xavier Mol
06/05/2026, 11:20
Shawn McKee
06/05/2026, 14:40
Eduardo Bach
06/05/2026, 15:20
Emily Kooistra,
Emily Kooistra
(NIKHEF),
Eric Kooistra
(ASTRON)
06/05/2026, 15:40
At Nikhef we run "fat" pool nodes, with each 100G uplink and 3 HW raidcards and around 1PB of net storage. Getting the most out of this requires careful tuning on the Linux size of things. I will walk you trough all the changes we have done and the rational why, and future improvements that can be made to achieve even better performance. Both on the node design and software changes in dcache.
