Timeouts increased to values which worked on Friday (local patch for testing). Once merged I will try to trigger it manually on another day to see the results in cdash without spoiling today's ones.
Probably further patches coming later as last week I could not test the L1 reco (some fixes not present in last week weekly tag)
The config files/scripts for these weekly build are in /lustre/cbm/users/ploizeau/cdash, this is also were I copy the log files and macro folder (minus all core_dumps an all_XXX.par) in one folder per weekly build. From what I saw with 2 builds, it takes around 9 GB per week which would mean around 500 GB per year
If this process matches what Iliya was asking, we could easily change this folder to somewhere in prod.