Hmm, that’s weird (kernel 2.6.33)
OK, i’m now convinced that Phonorix *does* know what they are doing with their kernel farms.
After upgrading to .33, the EXT-4 performance dropped so that I even understood that during… booting the system.
I really don’t understand why that is happening. Is EXT-4’s performance barriers so “dangerously” modified on the kernel, that we have to lose performance (and some system responsiveness) in the expense of security ? Because we need both.
Changing ext4 system attributes is dangerous as well ; barrier=0 is not so much of an option without battery or UPS, the fear of data loss is there and waiting.
I’m wondering if we do need ext4 after all, since it’s getting as fast as ext3 was, less reliable, with more bugs. Is it just … ext3.1 ? Was the speed of .29’s kernel just a bubble that burst ?
BtrFS is waiting patiently.
Powered by Blogilo