FWIW, I am looking at the munin graphs from the savannah nodes nearly all-day, every day. I see no obvious capacity problems (and I believe we can increase resources to the VCS2/3 nodes, if that would be required, but I don't believe that is a blocker).
I think Jing correctly reports the problem: the patches that have been attempted have no longer been applicable by the time the discussions required to develop them were completed. As a for-instance, Ayushman full completed this feature to the satisfaction of FSF tech staff (as well as Bob and I); however, between discussing those patches on the list and getting them perfected many thousands of lines of changes were committed to the repo and applied to the live instance rendering those patches invalid. It may be worth noting that those thousands of lines of changes, unlike Anush's work, were not discussed on this list. We -the other vols- saw them first when they were committed. In short: this is a "team discipline" problem, not anything technical. Here is the thread I'm referring to in my example given: https://lists.gnu.org/archive/html/savannah-hackers-public/2024-01/msg00018.html Thanks for taking an interest. Corwin On Mon, Oct 28, 2024 at 3:38 PM Alfred M. Szmidt <a...@gnu.org> wrote: > > This is a short sighted idea, waiting for things to magically get > fixed by next year. There have been too many outtages over the course > of this year, to the point where Savannah and Fencepost and other > stuff has been down for multiple days. > > CCing GAC and Zoe since something needs to be done _this_year_. From > the outside it looks like nothing is being done nor will be done, it > would be good if the FSF states what they are going to do to stabilize > our current infrastructure, cause this year hasn't been fun. >
