Xen
2016-04-27 10:50:13 UTC
1. Does LVM cache support discards of the underlying blocks (in the
cache) when the filesystem discards the blocks?
I was reading https://lwn.net/Articles/293658/ which makes it clear that
years ago kernel developers were introducing discard behaviour into
Linux filesystems with respect to flash devices and their need to copy
for wear leveling.
I know so little about it, but I have seen the "discard" flag mentioned
so much with respect to SSDs, that I must assume these discards are
there. Are LVM cache blocks discarded when the filesystem layer discards
these blocks?
Where can I find this info?
In https://www.redhat.com/archives/linux-lvm/2016-April/msg00030.html I
mentioned that such a discard feature would be necessary in order for a
filesystem to communicate to a block device layer which blocks are in
use, and for a block device layer to communicate back a set of available
blocks if these dynamically change.
I forgot the other question lol.
I'm interested in this solution to
"https://bugzilla.redhat.com/show_bug.cgi?id=1189215" but I will respond
in my other email (LVM Thin: Handle out of space conditions better).
cache) when the filesystem discards the blocks?
I was reading https://lwn.net/Articles/293658/ which makes it clear that
years ago kernel developers were introducing discard behaviour into
Linux filesystems with respect to flash devices and their need to copy
for wear leveling.
I know so little about it, but I have seen the "discard" flag mentioned
so much with respect to SSDs, that I must assume these discards are
there. Are LVM cache blocks discarded when the filesystem layer discards
these blocks?
Where can I find this info?
In https://www.redhat.com/archives/linux-lvm/2016-April/msg00030.html I
mentioned that such a discard feature would be necessary in order for a
filesystem to communicate to a block device layer which blocks are in
use, and for a block device layer to communicate back a set of available
blocks if these dynamically change.
I forgot the other question lol.
I'm interested in this solution to
"https://bugzilla.redhat.com/show_bug.cgi?id=1189215" but I will respond
in my other email (LVM Thin: Handle out of space conditions better).