btrfs: be less verbose on automatic bg reclaim
cover btrfs: be less verbose on automatic bg reclaim
From: Johannes Thumshirn <[email protected]>
BTRFS filesystems with active automatic block-group reclaim (this especially hits zoned file systems where automatic block-group reclaim is used for garbage collection) do a lot of log spamming, because every relocated block group is accompanied by three prints at info level.
The first patch removes the info message that is only present with automatic block group reclaim, we have a tracepoint right next to it so there's no need for the message at all.
The second patch introduces a verbose parameter for btrfs_relocate_chunk() and btrfs_relocate_block_group() to control if we want to add printks or not. Automatic reclaim calls into btrfs_relocate_chunk() setting verbose to false while the user-space triggered balance code path sets verbose to true retaining the old behaviour.
Johannes Thumshirn (2):
btrfs: remove redundant auto reclaim log message
btrfs: don't print relocation messages from auto reclaim
fs/btrfs/block-group.c | 8 +------- fs/btrfs/relocation.c | 12 ++++++++---- fs/btrfs/relocation.h | 3 ++- fs/btrfs/volumes.c | 14 ++++++++------ fs/btrfs/volumes.h | 3 ++- 5 files changed, 21 insertions(+), 19 deletions(-)
Signature · 1 line
2.50.0
PATCH 1/2 btrfs: remove redundant auto reclaim log message
From: Johannes Thumshirn <[email protected]>
Remove the log message before reclaiming a chunk in btrfs_reclaim_bgs_work(). Especially with automatic block-group reclaiming these messages spam the kernel log.
Note there is also a tracepoint for the same condition to ease debugging.
--- fs/btrfs/block-group.c | 6 ------ 1 file changed, 6 deletions(-)
fs/btrfs/block-group.c +0 -6
@@ -1963,12 +1963,6 @@ void btrfs_reclaim_bgs_work(struct work_struct *work) reserved = bg->reserved;spin_unlock (&bg->lock); -btrfs_info (fs_info, -"reclaiming chunk %llu with %llu%% used %llu%% reserved %llu%% unusable" , - bg->start, -div64_u64 (used *100 , bg->length), -div64_u64 (reserved *100 , bg->length), -div64_u64 (zone_unusable *100 , bg->length));trace_btrfs_reclaim_block_group (bg); ret =btrfs_relocate_chunk (fs_info, bg->start);if (ret) {
Signature · 1 line
2.50.0
PATCH 2/2 btrfs: don't print relocation messages from auto reclaim
In reply to Johannes ThumshirnFrom: Johannes Thumshirn <[email protected]>
When BTRFS is doing automatic block-group reclaim, it is spamming the kernel log messages a lot.
Add a verbose parameter to btrfs_relocate_chunk() and btrfs_relocate_block_group() to control the verbosity of these log message. This way the old behaviour of printing log messages on a user-space initiated balance operation can be kept while excessive log spamming due to auto reclaim is mitigated.
--- fs/btrfs/block-group.c | 2 +- fs/btrfs/relocation.c | 12 ++++++++---- fs/btrfs/relocation.h | 3 ++- fs/btrfs/volumes.c | 14 ++++++++------ fs/btrfs/volumes.h | 3 ++- 5 files changed, 21 insertions(+), 13 deletions(-)
fs/btrfs/block-group.c +1 -1
@@ -1964,7 +1964,7 @@ void btrfs_reclaim_bgs_work(struct work_struct *work)spin_unlock (&bg->lock);trace_btrfs_reclaim_block_group (bg); - ret =btrfs_relocate_chunk (fs_info, bg->start); + ret =btrfs_relocate_chunk (fs_info, bg->start, false);if (ret) {btrfs_dec_block_group_ro (bg); btrfs_err(fs_info, "error relocating chunk %llu",
fs/btrfs/relocation.c +8 -4
@@ -3908,7 +3908,8 @@ static const char *stage_to_string(enum reloc_stage stage)/* * function to relocate all extents in a block group. */ -int btrfs_relocate_block_group (struct btrfs_fs_info *fs_info,u64 group_start) +int btrfs_relocate_block_group (struct btrfs_fs_info *fs_info,u64 group_start, +bool verbose) {struct btrfs_block_group *bg;struct btrfs_root *extent_root =btrfs_extent_root (fs_info, group_start);
@@ -4000,7 +4001,8 @@ int btrfs_relocate_block_group(struct btrfs_fs_info *fs_info, u64 group_start) goto out; } -describe_relocation (rc->block_group); +if (verbose) +describe_relocation (rc->block_group);btrfs_wait_block_group_reservations (rc->block_group);btrfs_wait_nocow_writers (rc->block_group);
@@ -4044,8 +4046,10 @@ int btrfs_relocate_block_group(struct btrfs_fs_info *fs_info, u64 group_start)if (rc->extents_found ==0 )break ; -btrfs_info (fs_info,"found %llu extents, stage: %s" , - rc->extents_found,stage_to_string (finishes_stage)); +if (verbose) +btrfs_info (fs_info,"found %llu extents, stage: %s" , + rc->extents_found, +stage_to_string (finishes_stage)); }WARN_ON (rc->block_group->pinned >0 );
fs/btrfs/relocation.h +2 -1
@@ -12,7 +12,8 @@ struct btrfs_trans_handle;struct btrfs_ordered_extent ;struct btrfs_pending_snapshot ; -int btrfs_relocate_block_group (struct btrfs_fs_info *fs_info,u64 group_start); +int btrfs_relocate_block_group (struct btrfs_fs_info *fs_info,u64 group_start, +bool verbose);int btrfs_init_reloc_root (struct btrfs_trans_handle *trans,struct btrfs_root *root);int btrfs_update_reloc_root (struct btrfs_trans_handle *trans,struct btrfs_root *root);
fs/btrfs/volumes.c +8 -6
@@ -3412,7 +3412,8 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, u64 chunk_offset)return ret; } -int btrfs_relocate_chunk (struct btrfs_fs_info *fs_info,u64 chunk_offset) +int btrfs_relocate_chunk (struct btrfs_fs_info *fs_info,u64 chunk_offset, +bool verbose) {struct btrfs_root *root = fs_info->chunk_root;struct btrfs_trans_handle *trans;
@@ -3442,7 +3443,7 @@ int btrfs_relocate_chunk(struct btrfs_fs_info *fs_info, u64 chunk_offset)/* step one, relocate all the extents inside this chunk */ btrfs_scrub_pause (fs_info); - ret =btrfs_relocate_block_group (fs_info, chunk_offset); + ret =btrfs_relocate_block_group (fs_info, chunk_offset, true);btrfs_scrub_continue (fs_info);if (ret) {/*
@@ -3552,7 +3553,8 @@ static int btrfs_relocate_sys_chunks(struct btrfs_fs_info *fs_info)btrfs_release_path (path);if (chunk_type &BTRFS_BLOCK_GROUP_SYSTEM ) { - ret =btrfs_relocate_chunk (fs_info, found_key.offset); + ret =btrfs_relocate_chunk (fs_info, found_key.offset, + true);if (ret == -ENOSPC ) failed++;else
@@ -4217,7 +4219,7 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info) } } - ret =btrfs_relocate_chunk (fs_info, found_key.offset); + ret =btrfs_relocate_chunk (fs_info, found_key.offset, true);mutex_unlock (&fs_info->reclaim_bgs_lock);if (ret == -ENOSPC ) { enospc_errors++;
@@ -4985,7 +4987,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size) goto done; } - ret =btrfs_relocate_chunk (fs_info, chunk_offset); + ret =btrfs_relocate_chunk (fs_info, chunk_offset, true);mutex_unlock (&fs_info->reclaim_bgs_lock);if (ret == -ENOSPC ) { failed++;
@@ -8198,7 +8200,7 @@ static int relocating_repair_kthread(void *data)btrfs_info (fs_info,"zoned: relocating block group %llu to repair IO failure" , target); - ret =btrfs_relocate_chunk (fs_info, target); + ret =btrfs_relocate_chunk (fs_info, target, true); out:if (cache)
fs/btrfs/volumes.h +2 -1
@@ -765,7 +765,8 @@ void btrfs_describe_block_groups(u64 flags, char *buf, u32 size_buf);int btrfs_resume_balance_async (struct btrfs_fs_info *fs_info);int btrfs_recover_balance (struct btrfs_fs_info *fs_info);int btrfs_pause_balance (struct btrfs_fs_info *fs_info); -int btrfs_relocate_chunk (struct btrfs_fs_info *fs_info,u64 chunk_offset); +int btrfs_relocate_chunk (struct btrfs_fs_info *fs_info,u64 chunk_offset, +bool verbose);int btrfs_cancel_balance (struct btrfs_fs_info *fs_info);bool btrfs_chunk_writeable (struct btrfs_fs_info *fs_info,u64 chunk_offset);void btrfs_dev_stat_inc_and_print (struct btrfs_device *dev,int index);
Signature · 1 line
2.50.0
Quoted 11 lines · Johannes Thumshirn
From: Johannes Thumshirn <[email protected]>
BTRFS filesystems with active automatic block-group reclaim (this especially hits zoned file systems where automatic block-group reclaim is used for garbage collection) do a lot of log spamming, because every relocated block group is accompanied by three prints at info level.
The first patch removes the info message that is only present with automatic block group reclaim, we have a tracepoint right next to it so there's no need for the message at all.
Quoted 6 lines
The second patch introduces a
verboseparameter forbtrfs_relocate_chunk()andbtrfs_relocate_block_group()to control if we want to add printks or not. Automatic reclaim calls intobtrfs_relocate_chunk()settingverboseto false while the user-space triggered balance code path setsverboseto true retaining the old behaviour.
We also struggle with the spam at Meta with automatic reclaim enabled though quite a bit less with dynamic.. :) In particular, users often think it means there is some kind of btrfs error happening.
This looks like a good compromise to me, and I'd be quite happy to see it go in.
Thanks!
Quoted 15 lines
Johannes Thumshirn (2):
btrfs: remove redundant auto reclaim log message
btrfs: don't print relocation messages from auto reclaimfs/btrfs/block-group.c | 8 +------- fs/btrfs/relocation.c | 12 ++++++++---- fs/btrfs/relocation.h | 3 ++- fs/btrfs/volumes.c | 14 ++++++++------ fs/btrfs/volumes.h | 3 ++- 5 files changed, 21 insertions(+), 19 deletions(-)Signature · 1 line
2.50.0