Commit 97e728d4 authored by Josef Bacik's avatar Josef Bacik Committed by Chris Mason
Browse files

Btrfs: try to keep a healthy ratio of metadata vs data block groups 



This patch makes the chunk allocator keep a good ratio of metadata vs data
block groups.  By default for every 8 data block groups, we'll allocate 1
metadata chunk, or about 12% of the disk will be allocated for metadata.  This
can be changed by specifying the metadata_ratio mount option.

This is simply the number of data block groups that have to be allocated to
force a metadata chunk allocation.  By making sure we allocate metadata chunks
more often, we are less likely to get into situations where the whole disk
has been allocated as data block groups.

Signed-off-by: default avatarJosef Bacik <jbacik@redhat.com>
Signed-off-by: default avatarChris Mason <chris.mason@oracle.com>
parent 546888da

fs/btrfs/ctree.h

+3 −0
Original line number Diff line number Diff line
@@ -881,6 +881,9 @@ struct btrfs_fs_info { 
	u64 metadata_alloc_profile;

	u64 system_alloc_profile;



	unsigned data_chunk_allocations;

	unsigned metadata_ratio;



	void *bdev_holder;

};

fs/btrfs/disk-io.c

+1 −0
Original line number Diff line number Diff line
@@ -1604,6 +1604,7 @@ struct btrfs_root *open_ctree(struct super_block *sb, 
	fs_info->btree_inode = new_inode(sb);

	fs_info->btree_inode->i_ino = 1;

	fs_info->btree_inode->i_nlink = 1;

	fs_info->metadata_ratio = 8;



	fs_info->thread_pool_size = min_t(unsigned long,

					  num_online_cpus() + 2, 8);

fs/btrfs/extent-tree.c

+27 −1
Original line number Diff line number Diff line
@@ -1918,15 +1918,29 @@ void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode, 
	spin_unlock(&info->lock);

}



static void force_metadata_allocation(struct btrfs_fs_info *info)

{

	struct list_head *head = &info->space_info;

	struct btrfs_space_info *found;



	rcu_read_lock();

	list_for_each_entry_rcu(found, head, list) {

		if (found->flags & BTRFS_BLOCK_GROUP_METADATA)

			found->force_alloc = 1;

	}

	rcu_read_unlock();

}



static int do_chunk_alloc(struct btrfs_trans_handle *trans,

			  struct btrfs_root *extent_root, u64 alloc_bytes,

			  u64 flags, int force)

{

	struct btrfs_space_info *space_info;

	struct btrfs_fs_info *fs_info = extent_root->fs_info;

	u64 thresh;

	int ret = 0;



	mutex_lock(&extent_root->fs_info->chunk_mutex);

	mutex_lock(&fs_info->chunk_mutex);



	flags = btrfs_reduce_alloc_profile(extent_root, flags);
@@ -1958,6 +1972,18 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans, 
	}

	spin_unlock(&space_info->lock);



	/*

	 * if we're doing a data chunk, go ahead and make sure that

	 * we keep a reasonable number of metadata chunks allocated in the

	 * FS as well.

	 */

	if (flags & BTRFS_BLOCK_GROUP_DATA) {

		fs_info->data_chunk_allocations++;

		if (!(fs_info->data_chunk_allocations %

		      fs_info->metadata_ratio))

			force_metadata_allocation(fs_info);

	}



	ret = btrfs_alloc_chunk(trans, extent_root, flags);

	if (ret)

		space_info->full = 1;

fs/btrfs/super.c

+11 −1
Original line number Diff line number Diff line
@@ -68,7 +68,7 @@ enum { 
	Opt_degraded, Opt_subvol, Opt_device, Opt_nodatasum, Opt_nodatacow,

	Opt_max_extent, Opt_max_inline, Opt_alloc_start, Opt_nobarrier,

	Opt_ssd, Opt_thread_pool, Opt_noacl,  Opt_compress, Opt_notreelog,

	Opt_flushoncommit, Opt_err,

	Opt_ratio, Opt_flushoncommit, Opt_err,

};



static match_table_t tokens = {
@@ -87,6 +87,7 @@ static match_table_t tokens = { 
	{Opt_noacl, "noacl"},

	{Opt_notreelog, "notreelog"},

	{Opt_flushoncommit, "flushoncommit"},

	{Opt_ratio, "metadata_ratio=%d"},

	{Opt_err, NULL},

};
@@ -234,6 +235,15 @@ int btrfs_parse_options(struct btrfs_root *root, char *options) 
			printk(KERN_INFO "btrfs: turning on flush-on-commit\n");

			btrfs_set_opt(info->mount_opt, FLUSHONCOMMIT);

			break;

		case Opt_ratio:

			intarg = 0;

			match_int(&args[0], &intarg);

			if (intarg) {

				info->metadata_ratio = intarg;

				printk(KERN_INFO "btrfs: metadata ratio %d\n",

				       info->metadata_ratio);

			}

			break;

		default:

			break;

		}

CRA Git | Maintained and supported by SUSTech CRA and CCSE