Fred noticed mystery hepscore value in facility spreadsheet of 17.48 for R6625 with AMD 9354
Not clear where that value comes from
UM added 12 R6625s in 2024Q3 and MSU added 5 nodes 2025Q3
Measured hepscore is 22.57 (4 runs each on 3 msu and 2 um WNs)
Updating spreadsheet changed AGLT2 average 13.91 -> 14.54 i.e. +4.53%
Updated capacity tabs for 26Q1, 25Q4, 25Q3 but not able to change older tabs
Updated atlas CRIC and OSG topology
todo: submit update for hepix table
Also testing effect of vulnerability mitigation and updated CPU microcode
AMD 9354: 2.1% degradation in hepscore from mitigations and new microcode
Now testing some older CPU
Finished reconfiguring storage raid6 pools
which had default 64k stripe size instead of desired 512k for better IO performance
Status of Fireflies/SciTags in dCache
11.2.3-1 has almost all patches submitted by Shawn; functional but...
TPC transfers are missing the start marker which would specify the activity type
Transfers are still recorded via end markers but use default activity
non-TPC transfers don't specify activity and use default anyway
Added Kafka support in our dcache
with a kafka cluster on our 3 zoo nodes
early problems with flooding of the log partitions,
fixed it by reconfiguring it to limit the log size to 20GB