On Wed, 14 Mar 2012, Shaohua Li wrote:
Is it possible if you can directly build a fs in such ssd (that is not using raid), and check the performance of discard? I'd like to make sure it's not the problem of the ssd itself.
Yes, I just did the test on a partition without raid on the same SSD. Then there is hardly any difference if using discard or not. So it is not the problem of the SSD. It must be hanging somewhere in kernel. I once had an strace running on a rm command that was deleting thousand of small files and even the output of the strace command was stuck: 18:31:19.776588 unlinkat(7, "2K-3-4-16--207", 0) = 0 18:31:19.776699 unlinkat(7, "2K-3-4-16--44", 0) = 0 18:31:19.776808 unlinkat(7, "2K-3-4-16--1476", 0) = 0 18:31:19.776919 unlinkat(7, "2K-3-4-16--1277", 0 And when it continued after several minutes being frozen, the time given by strace was still 18:31:19.xxxx as if it did not hang. These hang occur when deleting lots of small files. But they not only effect the process doing the deleting, other process are also effected. When I tried to open a file on a tmpfs disk in another shell it too was stuck for several minutes. Regards, Holger -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
