Single process will always select the same mirror (raid1 and raid10),
selection si based on PID. Use a simple incremental counter on every
attempt. This is just a heuristic, we do not know which mirror device is
least loaded, but there is a good chance the distribution over mirrors
will improve.
Signed-off-by: David Sterba <dsterba@suse.cz>
---
fs/btrfs/volumes.c | 5 +++--
1 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index e138af7..3f0cc12 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -2893,6 +2893,7 @@ static int __btrfs_map_block(struct btrfs_mapping_tree
*map_tree, int rw,
int num_stripes;
int max_errors = 0;
struct btrfs_multi_bio *multi = NULL;
+ static unsigned int mirror_rotor = 0;
if (multi_ret && !(rw & (REQ_WRITE | REQ_DISCARD)))
stripes_allocated = 1;
@@ -3000,7 +3001,7 @@ again:
else {
stripe_index = find_live_mirror(map, 0,
map->num_stripes,
- current->pid % map->num_stripes);
+ mirror_rotor++ % map->num_stripes);
}
} else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
@@ -3026,7 +3027,7 @@ again:
else {
stripe_index = find_live_mirror(map, stripe_index,
map->sub_stripes, stripe_index +
- current->pid % map->sub_stripes);
+ mirror_rotor++ % map->sub_stripes);
}
} else {
/*
--
1.7.6
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html