Credits

[Michael H.W. Weber]

All RACs should go down at present due to the longer run times of individual WUs. RAC per se does not reflect computational throughput when WU runtimes are longer than RAC observation time window.
Example: When the RAC counter observes one week of result production but all WUs have runtimes of two weeks, within that week of observation, you have a RAC of zero. Still, your machines compute like mad. So, it is alright.
Also, as you can see our tweaks on the server side now (finally) start to grip. Even CMS WUs on hold on the server slowly move up. I hope this trend keeps up for a while.

Michael.

[yoyo]

Some days ago the transitioner and validator were not up to date. They were up to 20h behind. I think this was the reason for lower RAC.
yoyo