Hello, At Tuesday 05 February 2013 20:24:03 DaB. wrote:
Focusing solely on the replicated databases for a moment, they've been corrupt for months. Months and months. What can be done about this? Would having additional roots help with this?
it took a long time for different reasons (political ones, hardware-shortages, technology changes, learning and (for a very small part) personal-shortage). But the situation improved during the last weeks: I have written a replacement for trainwreck, the WMF provides dumps now on short-time-base (few days), we have a contact person there, I have learned how to import the new dump-format, and Nosy and I have setup 1 of the new servers as db-server. The s2-import will be the first step to replace/re-import all clusters. I'm not sure about the sequence yet (but s5 will be the next), but if everything works all clusters SHOULD be new withing 6 weeks. And no, in this special field another root would not help in my eyes.
Sincerely, DaB.