Not user support, but it’s feedback on the 4.2 version.
While testing 4.2, I accidentally launched a qube while out of memory (the xl info type memory, not dom0 memory) and the “steal memory from other qubes and give it to the new one” seems to actually work in 4.2
In the past versions I had never seen this work in a useful way. (I.E. on previous versions of qubes the newly launched qube would end up aborting)
This should make a major difference in how much memory is needed to run qubes (in a positive way)
What kind of system is qubes using to do that? Is there any way to detect how much memory has been “stolen” from VMs at a particular point in time?