Comment 29 for bug 913615

Dunno how much the Fedora kernel version can be trusted these days, but testing with a sampling of kernels grabbed from koji, this doesn't seem to start at 2.6.39 but 2.6.40 and gradually getting worse. This is the same test-case as with the above numbers, 3.2.3 retested for curiosity and producing a little less horrible figure this time, but still by far slowest of them all:

[root@localhost slowdb]# uname -r
2.6.38.6-26.rc1.fc15.x86_64
[root@localhost slowdb]# time rpmdb --initdb --dbpath /tmp/slowdb/
warning: Generating 12 missing index(es), please wait...

real 2m39.459s
user 0m6.173s
sys 0m2.478s

[root@localhost slowdb]# uname -r
2.6.39-1.fc16.x86_64
[root@localhost slowdb]# time rpmdb --initdb --dbpath /tmp/slowdb/
warning: Generating 12 missing index(es), please wait...

real 2m24.814s
user 0m5.939s
sys 0m2.245s

[root@localhost slowdb]# uname -r
2.6.40-4.fc15.x86_64
[root@localhost slowdb]# time rpmdb --initdb --dbpath /tmp/slowdb/
warning: Generating 12 missing index(es), please wait...

real 6m26.406s
user 0m5.772s
sys 0m2.152s

[root@localhost slowdb]# uname -r
2.6.41.10-3.fc15.x86_64
[root@localhost slowdb]# time rpmdb --initdb --dbpath /tmp/slowdb/
warning: Generating 12 missing index(es), please wait...

real 5m42.944s
user 0m5.796s
sys 0m2.164s

[root@localhost slowdb]# uname -r
3.2.3-2.fc16.x86_64
[root@localhost slowdb]# time rpmdb --initdb --dbpath /tmp/slowdb/
warning: Generating 12 missing index(es), please wait...

real 7m53.031s
user 0m5.750s
sys 0m2.217s