Hi all, This patch series add inode steal mechanism for inode allocation. Modification from V2 to V3: 1. Add a new member to record the times we have stealed inode from other slots so that we can go directly to inode steal without trying our own slot every time. In OCFS2, we allocate the inodes from slot specific inode_alloc to avoid inode creation congestion. The local alloc file grows in a large contiguous chunk. As for a 4K bs, it grows 4M every time. So 1024 inodes will be allocated at a time. Over time, if the fs gets fragmented enough(e.g, the user has created many small files and also delete some of them), we can end up in a situation, whereby we cannot extend the inode_alloc as we don't have a large chunk free in the global_bitmap even if df shows few gigs free. More annoying is that this situation will invariably mean that while one cannot create inodes on one node but can from another node. Still more annoying is that an unused slot may have space for plenty of inodes but is unusable as the user may not be mounting as many nodes anymore. This patch series implement a solution which is to steal inodes from another slot. 2 new variables are added for it. They are: 1) ocfs2_super->inode_steal_slot. It is initalized as invalid and only set valid when we steal inode from other slots successfully. When we flush the truncate log, complete local alloc recovery or allocate from our own slot successfully, it will be reset to invalid. 2) inode_steal_times. It is used to record the times we try to steal inode from other nodes. And it is increased no matter whether our steal succeed or not. It is reset to zero when we try to allocate from our own slot. So with this 2 new variables, now the whole inode allocation process is: 1. Check whether the ocfs2_super->inode_steal_slot is valid. If it is invalid, goto step 2, that is to try to allocate from our own. If it is valid, then we must have stealed inode successfully just now, so verify whether we have steal "inode_steal_times". If yes, goto step 2 since now we need to try own slot in case there is some space for us. If not, goto step 3 and steal from other nodes directly. 2. Allocate from its own inode_alloc:000X and zero inode_steal_times. 1) If we can reserve, OK. 2) If fails, try to allocate a large chunk and reserve once again. 3) If OK, clear ocfs2_super->inode_steal_slot and exit directly. 3. Try to allocate from other nodes. 1) If ocfs2_super->inode_steal_slot is valid, start from that node, otherwise start from the node next to us. This time, Just try to reserve in inode_alloc, we don't go for global_bitmap if this node also can't allocate the inode. 3) Try the node next until we reach the first steal slot again. 4) If we succeed in one node's inode_alloc, set ocfs2_super->inode_steal_slot to it. 5) increase inode_steal_times.
Tao Ma
2008-Mar-03 01:13 UTC
[Ocfs2-devel] [PATCH 1/3] Add a new parameter for ocfs2_reserve_suballoc_bits.V3
In some cases(Inode stealing from other nodes), we may not want ocfs2_reserve_suballoc_bits to allocate new groups from the global_bitmap since it may already be full. So add a new parameter for this. Signed-off-by: Tao Ma <tao.ma@oracle.com> Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com> --- fs/ocfs2/suballoc.c | 22 ++++++++++++++++++---- 1 files changed, 18 insertions(+), 4 deletions(-) diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 72c198a..3be4e73 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -46,6 +46,9 @@ #include "buffer_head_io.h" +#define NOT_ALLOC_NEW_GROUP 0 +#define ALLOC_NEW_GROUP 1 + static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg); static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe); static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl); @@ -391,7 +394,8 @@ bail: static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb, struct ocfs2_alloc_context *ac, int type, - u32 slot) + u32 slot, + int alloc_new_group) { int status; u32 bits_wanted = ac->ac_bits_wanted; @@ -446,6 +450,14 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb, goto bail; } + if (alloc_new_group != ALLOC_NEW_GROUP) { + mlog(0, "Alloc File %u Full: wanted=%u, free_bits=%u, " + "and we don't alloc a new group for it.\n", + slot, bits_wanted, free_bits); + status = -ENOSPC; + goto bail; + } + status = ocfs2_block_group_alloc(osb, alloc_inode, bh); if (status < 0) { if (status != -ENOSPC) @@ -490,7 +502,8 @@ int ocfs2_reserve_new_metadata(struct ocfs2_super *osb, (*ac)->ac_group_search = ocfs2_block_group_search; status = ocfs2_reserve_suballoc_bits(osb, (*ac), - EXTENT_ALLOC_SYSTEM_INODE, slot); + EXTENT_ALLOC_SYSTEM_INODE, + slot, ALLOC_NEW_GROUP); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); @@ -527,7 +540,7 @@ int ocfs2_reserve_new_inode(struct ocfs2_super *osb, status = ocfs2_reserve_suballoc_bits(osb, *ac, INODE_ALLOC_SYSTEM_INODE, - osb->slot_num); + osb->slot_num, ALLOC_NEW_GROUP); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); @@ -557,7 +570,8 @@ int ocfs2_reserve_cluster_bitmap_bits(struct ocfs2_super *osb, status = ocfs2_reserve_suballoc_bits(osb, ac, GLOBAL_BITMAP_SYSTEM_INODE, - OCFS2_INVALID_SLOT); + OCFS2_INVALID_SLOT, + ALLOC_NEW_GROUP); if (status < 0 && status != -ENOSPC) { mlog_errno(status); goto bail; -- 1.5.3.GIT
Tao Ma
2008-Mar-03 01:13 UTC
[Ocfs2-devel] [PATCH 2/3] Add ac_alloc_slot in ocfs2_alloc_context.V3
In inode stealing, we no longer restrict the allocation to happen in the local node. So it is neccessary for us to add a new member in ocfs2_alloc_context to indicate which slot we are using for allocation. We also modify the process of local alloc so that this member can be used there also. Signed-off-by: Tao Ma <tao.ma@oracle.com> --- fs/ocfs2/localalloc.c | 2 ++ fs/ocfs2/suballoc.c | 1 + fs/ocfs2/suballoc.h | 1 + 3 files changed, 4 insertions(+), 0 deletions(-) diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c index add1ffd..80d1c75 100644 --- a/fs/ocfs2/localalloc.c +++ b/fs/ocfs2/localalloc.c @@ -526,6 +526,8 @@ int ocfs2_reserve_local_alloc_bits(struct ocfs2_super *osb, } ac->ac_inode = local_alloc_inode; + /* We should never use localalloc from another slot */ + ac->ac_alloc_slot = osb->slot_num; ac->ac_which = OCFS2_AC_USE_LOCAL; get_bh(osb->local_alloc_bh); ac->ac_bh = osb->local_alloc_bh; diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 3be4e73..33d5573 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -424,6 +424,7 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb, } ac->ac_inode = alloc_inode; + ac->ac_alloc_slot = slot; fe = (struct ocfs2_dinode *) bh->b_data; if (!OCFS2_IS_VALID_DINODE(fe)) { diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h index 8799033..544c600 100644 --- a/fs/ocfs2/suballoc.h +++ b/fs/ocfs2/suballoc.h @@ -36,6 +36,7 @@ typedef int (group_search_t)(struct inode *, struct ocfs2_alloc_context { struct inode *ac_inode; /* which bitmap are we allocating from? */ struct buffer_head *ac_bh; /* file entry bh */ + u32 ac_alloc_slot; /* which slot are we allocating from? */ u32 ac_bits_wanted; u32 ac_bits_given; #define OCFS2_AC_USE_LOCAL 1 -- 1.5.3.GIT
Tao Ma
2008-Mar-03 01:13 UTC
[Ocfs2-devel] [PATCH 3/3] Add inode stealing for ocfs2_reserve_new_inode.V3
2 new variables are added. They are: 1) ocfs2_super->inode_steal_slot. It is initalized as invalid and only set valid when we steal inode from other slots successfully. When we flush the truncate log, complete local alloc recovery or allocate from our own slot successfully, it will be reset to invalid. 2) inode_steal_times. It is used to record the times we try to steal inode from other nodes. And it is increased no matter whether our steal succeed or not. It is reset to zero when we try to allocate from our own slot. So with this 2 new variables, now the whole inode allocation process is: 1. Check whether the ocfs2_super->inode_steal_slot is valid. If it is invalid, goto step 2, that is to try to allocate from our own. If it is valid, then we must have stealed inode successfully just now, so verify whether we have steal "inode_steal_times". If yes, goto step 2 since now we need to try own slot in case there is some space for us. If not, goto step 3 and steal from other nodes directly. 2. Allocate from its own inode_alloc:000X and zero inode_steal_times. 1) If we can reserve, OK. 2) If fails, try to allocate a large chunk and reserve once again. 3) If OK, clear ocfs2_super->inode_steal_slot and exit directly. 3. Try to allocate from other nodes. 1) If ocfs2_super->inode_steal_slot is valid, start from that node, otherwise start from the node next to us. This time, Just try to reserve in inode_alloc, we don't go for global_bitmap if this node also can't allocate the inode. 3) Try the node next until we reach the first steal slot again. 4) If we succeed in one node's inode_alloc, set ocfs2_super->inode_steal_slot to it. 5) increase inode_steal_times. Signed-off-by: Tao Ma <tao.ma@oracle.com> --- fs/ocfs2/alloc.c | 2 + fs/ocfs2/localalloc.c | 2 + fs/ocfs2/namei.c | 2 +- fs/ocfs2/ocfs2.h | 32 +++++++++++++++++- fs/ocfs2/suballoc.c | 83 +++++++++++++++++++++++++++++++++++++++++++++++- fs/ocfs2/super.c | 1 + 6 files changed, 117 insertions(+), 5 deletions(-) diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 447206e..f333cdc 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -4788,6 +4788,8 @@ static void ocfs2_truncate_log_worker(struct work_struct *work) status = ocfs2_flush_truncate_log(osb); if (status < 0) mlog_errno(status); + else + ocfs2_init_inode_steal_slot(osb); mlog_exit(status); } diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c index 80d1c75..ca3bf0c 100644 --- a/fs/ocfs2/localalloc.c +++ b/fs/ocfs2/localalloc.c @@ -450,6 +450,8 @@ out_mutex: iput(main_bm_inode); out: + if (!status) + ocfs2_init_inode_steal_slot(osb); mlog_exit(status); return status; } diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index ae9ad95..ab5a227 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -424,7 +424,7 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb, fe->i_fs_generation = cpu_to_le32(osb->fs_generation); fe->i_blkno = cpu_to_le64(fe_blkno); fe->i_suballoc_bit = cpu_to_le16(suballoc_bit); - fe->i_suballoc_slot = cpu_to_le16(osb->slot_num); + fe->i_suballoc_slot = cpu_to_le16(inode_ac->ac_alloc_slot); fe->i_uid = cpu_to_le32(current->fsuid); if (dir->i_mode & S_ISGID) { fe->i_gid = cpu_to_le32(dir->i_gid); diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 6546cef..2278ec0 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -206,11 +206,13 @@ struct ocfs2_super u32 s_feature_incompat; u32 s_feature_ro_compat; - /* Protects s_next_generaion, osb_flags. Could protect more on - * osb as it's very short lived. */ + /* Protects s_next_generation, osb_flags and s_inode_steal_slot. + * Could protect more on osb as it's very short lived. + */ spinlock_t osb_lock; u32 s_next_generation; unsigned long osb_flags; + s16 s_inode_steal_slot; unsigned long s_mount_opt; unsigned int s_atime_quantum; @@ -522,6 +524,32 @@ static inline unsigned int ocfs2_pages_per_cluster(struct super_block *sb) return pages_per_cluster; } +static inline void ocfs2_init_inode_steal_slot(struct ocfs2_super *osb) +{ + spin_lock(&osb->osb_lock); + osb->s_inode_steal_slot = OCFS2_INVALID_SLOT; + spin_unlock(&osb->osb_lock); +} + +static inline void ocfs2_set_inode_steal_slot(struct ocfs2_super *osb, + u16 slot) +{ + spin_lock(&osb->osb_lock); + osb->s_inode_steal_slot = slot; + spin_unlock(&osb->osb_lock); +} + +static inline u16 ocfs2_get_inode_steal_slot(struct ocfs2_super *osb) +{ + s16 slot; + + spin_lock(&osb->osb_lock); + slot = osb->s_inode_steal_slot; + spin_unlock(&osb->osb_lock); + + return slot; +} + #define ocfs2_set_bit ext2_set_bit #define ocfs2_clear_bit ext2_clear_bit #define ocfs2_test_bit ext2_test_bit diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 33d5573..dad4b70 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -109,7 +109,7 @@ static inline void ocfs2_block_to_cluster_group(struct inode *inode, u64 *bg_blkno, u16 *bg_bit_off); -void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac) +static void ocfs2_free_ac_resource(struct ocfs2_alloc_context *ac) { struct inode *inode = ac->ac_inode; @@ -120,9 +120,17 @@ void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac) mutex_unlock(&inode->i_mutex); iput(inode); + ac->ac_inode = NULL; } - if (ac->ac_bh) + if (ac->ac_bh) { brelse(ac->ac_bh); + ac->ac_bh = NULL; + } +} + +void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac) +{ + ocfs2_free_ac_resource(ac); kfree(ac); } @@ -522,10 +530,47 @@ bail: return status; } +static int ocfs2_steal_inode_from_other_nodes(struct ocfs2_super *osb, + struct ocfs2_alloc_context *ac) +{ + int status = -ENOSPC, i; + s16 slot = ocfs2_get_inode_steal_slot(osb); + + /* + * This is the first time we try to steal inodes, so begin + * with the node next to us first. + */ + if (slot == OCFS2_INVALID_SLOT) + slot = osb->slot_num + 1; + + for (i = 0; i < osb->max_slots; i++, slot++) { + if (slot == osb->max_slots) + slot = 0; + + if (slot == osb->slot_num) + continue; + + status = ocfs2_reserve_suballoc_bits(osb, ac, + INODE_ALLOC_SYSTEM_INODE, + slot, NOT_ALLOC_NEW_GROUP); + if (status >= 0) { + ocfs2_set_inode_steal_slot(osb, slot); + break; + } + + ocfs2_free_ac_resource(ac); + } + + return status; +} + int ocfs2_reserve_new_inode(struct ocfs2_super *osb, struct ocfs2_alloc_context **ac) { int status; + s16 slot = ocfs2_get_inode_steal_slot(osb); +#define OCFS2_INODE_STEAL_TIMES 10 + static atomic_t inode_steal_times = ATOMIC_INIT(0); *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL); if (!(*ac)) { @@ -539,9 +584,43 @@ int ocfs2_reserve_new_inode(struct ocfs2_super *osb, (*ac)->ac_group_search = ocfs2_block_group_search; + /* + * slot is set when we successfully steal inode from other nodes. + * It is reset in 3 places: + * 1. when we flush the turncate log + * 2. when we complete local alloc recovery. + * 3. when we successfully allocate from our own slot. + * After it is set, we will go on stealing inodes until we find the + * need to check our slots to see wether there is some space for us. + */ + if (slot != OCFS2_INVALID_SLOT && + atomic_read(&inode_steal_times) < OCFS2_INODE_STEAL_TIMES) + goto inode_steal; + + atomic_set(&inode_steal_times, 0); status = ocfs2_reserve_suballoc_bits(osb, *ac, INODE_ALLOC_SYSTEM_INODE, osb->slot_num, ALLOC_NEW_GROUP); + if (status >= 0) { + status = 0; + + /* + * Some inodes must be freed by us, so try to allocate + * from our own next time. + */ + if (slot != OCFS2_INVALID_SLOT) + ocfs2_init_inode_steal_slot(osb); + goto bail; + } else if (status < 0 && status != -ENOSPC) { + mlog_errno(status); + goto bail; + } + + ocfs2_free_ac_resource(*ac); + +inode_steal: + status = ocfs2_steal_inode_from_other_nodes(osb, *ac); + atomic_inc(&inode_steal_times); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index bec75af..c4e82c7 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1193,6 +1193,7 @@ static int ocfs2_mount_volume(struct super_block *sb) mlog_errno(status); goto leave; } + ocfs2_init_inode_steal_slot(osb); /* load all node-local system inodes */ status = ocfs2_init_local_system_inodes(osb); -- 1.5.3.GIT
Sunil Mushran
2008-Mar-03 18:25 UTC
[Ocfs2-devel] [PATCH 2/3] Add ac_alloc_slot in ocfs2_alloc_context.V3
Tao Ma wrote:> In inode stealing, we no longer restrict the allocation to > happen in the local node. So it is neccessary for us to add > a new member in ocfs2_alloc_context to indicate which slot > we are using for allocation. We also modify the process of > local alloc so that this member can be used there also. > > Signed-off-by: Tao Ma <tao.ma@oracle.com> >Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>> --- > fs/ocfs2/localalloc.c | 2 ++ > fs/ocfs2/suballoc.c | 1 + > fs/ocfs2/suballoc.h | 1 + > 3 files changed, 4 insertions(+), 0 deletions(-) > > diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c > index add1ffd..80d1c75 100644 > --- a/fs/ocfs2/localalloc.c > +++ b/fs/ocfs2/localalloc.c > @@ -526,6 +526,8 @@ int ocfs2_reserve_local_alloc_bits(struct ocfs2_super *osb, > } > > ac->ac_inode = local_alloc_inode; > + /* We should never use localalloc from another slot */ > + ac->ac_alloc_slot = osb->slot_num; > ac->ac_which = OCFS2_AC_USE_LOCAL; > get_bh(osb->local_alloc_bh); > ac->ac_bh = osb->local_alloc_bh; > diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c > index 3be4e73..33d5573 100644 > --- a/fs/ocfs2/suballoc.c > +++ b/fs/ocfs2/suballoc.c > @@ -424,6 +424,7 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb, > } > > ac->ac_inode = alloc_inode; > + ac->ac_alloc_slot = slot; > > fe = (struct ocfs2_dinode *) bh->b_data; > if (!OCFS2_IS_VALID_DINODE(fe)) { > diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h > index 8799033..544c600 100644 > --- a/fs/ocfs2/suballoc.h > +++ b/fs/ocfs2/suballoc.h > @@ -36,6 +36,7 @@ typedef int (group_search_t)(struct inode *, > struct ocfs2_alloc_context { > struct inode *ac_inode; /* which bitmap are we allocating from? */ > struct buffer_head *ac_bh; /* file entry bh */ > + u32 ac_alloc_slot; /* which slot are we allocating from? */ > u32 ac_bits_wanted; > u32 ac_bits_given; > #define OCFS2_AC_USE_LOCAL 1 >