Comment 22 for bug 913615

Seeing this: http://lwn.net/Articles/478440/ (original report at http://article.gmane.org/gmane.linux.kernel/1244680 for those who can't yet access the LWN article) makes me wonder if that's what is hitting rpm.

Just a theory for now, but at any rate that's the kind of thing that would explain why the very same DB configuration worked fine in (early) F15 but things got dirt slow in F16: F15 started with kernel 2.6.38 (but later updates brought in 3.0.x masquerading as 2.6.40 iirc) whereas F16 had 3.x from day one. If somebody wants to try and recompile their kernel to see if it makes any difference, this is what's supposed to fix it: http://article.gmane.org/gmane.linux.kernel.mm/73293