Skip to contentSkip to conversation

Patches that adds delayed orphan scan timer (rev 3)

Ocfs2-devel New Srinivas Eeda 4 participants 11 messages Jun 4, 2009 – Jul 17, 2009 OCFS2

lore
Share & tools
lore
#1 lore raw

PATCH 1/2 ocfs2: timer to queue scan of all orphan slots

When a dentry is unlinked, the unlinking node takes an EX on the dentry lock before moving the dentry to the orphan directory. The other nodes, that all had a PR on the same dentry lock, flag the corresponding inode as MAYBE_ORPHANED during the downconvert. The inode is finally deleted when the last node to iput the inode notices the MAYBE_ORPHANED flag.

A problem arises if a node is forced to free dentry locks because of memory pressure. If this happens, the node will no longer get downconvert notifications for the dentries that have been unlinked on another node. If it also happens that node is actively using the corresponding inode and happens to be the one performing the last iput on that inode, it will fail to delete the inode as it will not have the MAYBE_ORPHANED flag set.

This patch fixes this shortcoming by introducing a periodic scan of the orphan directories to delete such inodes. Care has been taken to distribute the workload across the cluster so that no one node has to perform the task all the time.

---
 fs/ocfs2/dlmglue.c      |   51 ++++++++++++++++++++++
 fs/ocfs2/dlmglue.h      |   11 +++++
 fs/ocfs2/journal.c      |  106 +++++++++++++++++++++++++++++++++++++++++++++++
 fs/ocfs2/journal.h      |    4 ++
 fs/ocfs2/ocfs2.h        |   10 ++++
 fs/ocfs2/ocfs2_lockid.h |    5 ++
 fs/ocfs2/super.c        |    9 ++++
 7 files changed, 196 insertions(+), 0 deletions(-)
fs/ocfs2/dlmglue.c +51 -0
index e15fc7d..0f35b83 100644
@@ -248,6 +248,10 @@ static struct ocfs2_lock_res_ops ocfs2_nfs_sync_lops = {
 	.flags		= 0,
 };
 
+static struct ocfs2_lock_res_ops ocfs2_orphan_scan_lops = {
+	.flags		= LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
+};
+
 static struct ocfs2_lock_res_ops ocfs2_dentry_lops = {
 	.get_osb	= ocfs2_get_dentry_osb,
 	.post_unlock	= ocfs2_dentry_post_unlock,
@@ -637,6 +641,19 @@ static void ocfs2_nfs_sync_lock_res_init(struct ocfs2_lock_res *res,
 				   &ocfs2_nfs_sync_lops, osb);
 }
 
+static void ocfs2_orphan_scan_lock_res_init(struct ocfs2_lock_res *res,
+					    struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan_lvb *lvb;
+
+	ocfs2_lock_res_init_once(res);
+	ocfs2_build_lock_name(OCFS2_LOCK_TYPE_ORPHAN_SCAN, 0, 0, res->l_name);
+	ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_ORPHAN_SCAN,
+				   &ocfs2_orphan_scan_lops, osb);
+	lvb = ocfs2_dlm_lvb(&res->l_lksb);
+	lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+}
+
 void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
 			      struct ocfs2_file_private *fp)
 {
@@ -2352,6 +2369,37 @@ void ocfs2_inode_unlock(struct inode *inode,
 	mlog_exit_void();
 }
 
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex)
+{
+	struct ocfs2_lock_res *lockres;
+	struct ocfs2_orphan_scan_lvb *lvb;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+	int status = 0;
+
+	lockres = &osb->osb_orphan_scan.os_lockres;
+	status = ocfs2_cluster_lock(osb, lockres, level, 0, 0); 
+	if (status < 0)
+		return status;
+
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	if (lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
+		*seqno = be32_to_cpu(lvb->lvb_os_seqno); 
+	return status;
+}
+
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, int ex)
+{
+	struct ocfs2_lock_res *lockres;
+	struct ocfs2_orphan_scan_lvb *lvb;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+
+	lockres = &osb->osb_orphan_scan.os_lockres;
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+	lvb->lvb_os_seqno = cpu_to_be32(seqno);
+	ocfs2_cluster_unlock(osb, lockres, level); 
+}
+
 int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex)
 {
@@ -2842,6 +2890,7 @@ local:
 	ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
 	ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
 	ocfs2_nfs_sync_lock_res_init(&osb->osb_nfs_sync_lockres, osb);
+	ocfs2_orphan_scan_lock_res_init(&osb->osb_orphan_scan.os_lockres, osb);
 
 	osb->cconn = conn;
 
@@ -2878,6 +2927,7 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
 	ocfs2_lock_res_free(&osb->osb_super_lockres);
 	ocfs2_lock_res_free(&osb->osb_rename_lockres);
 	ocfs2_lock_res_free(&osb->osb_nfs_sync_lockres);
+	ocfs2_lock_res_free(&osb->osb_orphan_scan.os_lockres);
 
 	ocfs2_cluster_disconnect(osb->cconn, hangup_pending);
 	osb->cconn = NULL;
@@ -3061,6 +3111,7 @@ static void ocfs2_drop_osb_locks(struct ocfs2_super *osb)
 	ocfs2_simple_drop_lockres(osb, &osb->osb_super_lockres);
 	ocfs2_simple_drop_lockres(osb, &osb->osb_rename_lockres);
 	ocfs2_simple_drop_lockres(osb, &osb->osb_nfs_sync_lockres);
+	ocfs2_simple_drop_lockres(osb, &osb->osb_orphan_scan.os_lockres);
 }
 
 int ocfs2_drop_inode_locks(struct inode *inode)
fs/ocfs2/dlmglue.h +11 -0
index e1fd572..31b90d7 100644
@@ -62,6 +62,14 @@ struct ocfs2_qinfo_lvb {
 	__be32	lvb_free_entry;
 };
 
+#define OCFS2_ORPHAN_LVB_VERSION 1
+
+struct ocfs2_orphan_scan_lvb {
+	__u8	lvb_version;
+	__u8	lvb_reserved[3];
+	__be32	lvb_os_seqno;
+};
+
 /* ocfs2_inode_lock_full() 'arg_flags' flags */
 /* don't wait on recovery. */
 #define OCFS2_META_LOCK_RECOVERY	(0x01)
@@ -113,6 +121,9 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex);
 void ocfs2_super_unlock(struct ocfs2_super *osb,
 			int ex);
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex);
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, int ex);
+
 int ocfs2_rename_lock(struct ocfs2_super *osb);
 void ocfs2_rename_unlock(struct ocfs2_super *osb);
 int ocfs2_nfs_sync_lock(struct ocfs2_super *osb, int ex);
fs/ocfs2/journal.c +106 -0

124 lines not on this page · raw message

fs/ocfs2/journal.h +4 -0
index 619dd7f..3483202 100644
@@ -144,6 +144,10 @@ static inline void ocfs2_inode_set_new(struct ocfs2_super *osb,
 }
 
 /* Exported only for the journal struct init code in super.c. Do not call. */
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_exit(struct ocfs2_super *osb);
+
 void ocfs2_complete_recovery(struct work_struct *work);
 void ocfs2_wait_for_recovery(struct ocfs2_super *osb);
 
fs/ocfs2/ocfs2.h +10 -0
index 1386281..373fb1c 100644
@@ -151,6 +151,14 @@ struct ocfs2_lock_res {
 #endif
 };
 
+struct ocfs2_orphan_scan {
+	struct mutex 		os_lock;
+	struct ocfs2_super 	*os_osb;
+	struct ocfs2_lock_res 	os_lockres;     /* lock to synchronize scans */
+	struct delayed_work 	os_orphan_scan_work;
+	u32  			os_seqno;       /* incremented on every scan */
+};
+
 struct ocfs2_dlm_debug {
 	struct kref d_refcnt;
 	struct dentry *d_locking_state;
@@ -341,6 +349,8 @@ struct ocfs2_super
 	unsigned int			*osb_orphan_wipes;
 	wait_queue_head_t		osb_wipe_event;
 
+	struct ocfs2_orphan_scan 	osb_orphan_scan; 
+
 	/* used to protect metaecc calculation check of xattr. */
 	spinlock_t osb_xattr_lock;
 
fs/ocfs2/ocfs2_lockid.h +5 -0
index a53ce87..fcdba09 100644
@@ -48,6 +48,7 @@ enum ocfs2_lock_type {
 	OCFS2_LOCK_TYPE_FLOCK,
 	OCFS2_LOCK_TYPE_QINFO,
 	OCFS2_LOCK_TYPE_NFS_SYNC,
+	OCFS2_LOCK_TYPE_ORPHAN_SCAN,
 	OCFS2_NUM_LOCK_TYPES
 };
 
@@ -85,6 +86,9 @@ static inline char ocfs2_lock_type_char(enum ocfs2_lock_type type)
 		case OCFS2_LOCK_TYPE_NFS_SYNC:
 			c = 'Y';
 			break;
+		case OCFS2_LOCK_TYPE_ORPHAN_SCAN:
+			c = 'P';
+			break;
 		default:
 			c = '\0';
 	}
@@ -104,6 +108,7 @@ static char *ocfs2_lock_type_strings[] = {
 	[OCFS2_LOCK_TYPE_OPEN] = "Open",
 	[OCFS2_LOCK_TYPE_FLOCK] = "Flock",
 	[OCFS2_LOCK_TYPE_QINFO] = "Quota",
+	[OCFS2_LOCK_TYPE_ORPHAN_SCAN] = "OrphanScan",
 };
 
 static inline const char *ocfs2_lock_type_string(enum ocfs2_lock_type type)
fs/ocfs2/super.c +9 -0
index 79ff8d9..44ac27e 100644
@@ -1802,6 +1802,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 
 	ocfs2_truncate_log_shutdown(osb);
 
+	ocfs2_orphan_scan_stop(osb);
+
 	/* This will disable recovery and flush any recovery work. */
 	ocfs2_recovery_exit(osb);
 
@@ -1957,6 +1959,13 @@ static int ocfs2_initialize_super(struct super_block *sb,
 		goto bail;
 	}
 
+	status = ocfs2_orphan_scan_init(osb);
+	if (status) {
+		mlog(ML_ERROR, "Unable to initialize delayed orphan scan\n");
+		mlog_errno(status);
+		goto bail;
+	}
+
 	init_waitqueue_head(&osb->checkpoint_event);
 	atomic_set(&osb->needs_checkpoint, 0);
 
Signature · 1 line
1.5.6.5
#2 lore raw

PATCH 1/2 ocfs2: timer to queue scan of all orphan slots

Quoted 387 lines · Srinivas Eeda

When a dentry is unlinked, the unlinking node takes an EX on the dentry lock before moving the dentry to the orphan directory. The other nodes, that all had a PR on the same dentry lock, flag the corresponding inode as MAYBE_ORPHANED during the downconvert. The inode is finally deleted when the last node to iput the inode notices the MAYBE_ORPHANED flag.

A problem arises if a node is forced to free dentry locks because of memory pressure. If this happens, the node will no longer get downconvert notifications for the dentries that have been unlinked on another node. If it also happens that node is actively using the corresponding inode and happens to be the one performing the last iput on that inode, it will fail to delete the inode as it will not have the MAYBE_ORPHANED flag set.

This patch fixes this shortcoming by introducing a periodic scan of the orphan directories to delete such inodes. Care has been taken to distribute the workload across the cluster so that no one node has to perform the task all the time.

---
 fs/ocfs2/dlmglue.c      |   51 ++++++++++++++++++++++
 fs/ocfs2/dlmglue.h      |   11 +++++
 fs/ocfs2/journal.c      |  106 +++++++++++++++++++++++++++++++++++++++++++++++
 fs/ocfs2/journal.h      |    4 ++
 fs/ocfs2/ocfs2.h        |   10 ++++
 fs/ocfs2/ocfs2_lockid.h |    5 ++
 fs/ocfs2/super.c        |    9 ++++
 7 files changed, 196 insertions(+), 0 deletions(-)
fs/ocfs2/dlmglue.c +51 -0
index e15fc7d..0f35b83 100644
@@ -248,6 +248,10 @@ static struct ocfs2_lock_res_ops ocfs2_nfs_sync_lops = {
 	.flags		= 0,
 };
 
+static struct ocfs2_lock_res_ops ocfs2_orphan_scan_lops = {
+	.flags		= LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
+};
+
 static struct ocfs2_lock_res_ops ocfs2_dentry_lops = {
 	.get_osb	= ocfs2_get_dentry_osb,
 	.post_unlock	= ocfs2_dentry_post_unlock,
@@ -637,6 +641,19 @@ static void ocfs2_nfs_sync_lock_res_init(struct ocfs2_lock_res *res,
 				   &ocfs2_nfs_sync_lops, osb);
 }
 
+static void ocfs2_orphan_scan_lock_res_init(struct ocfs2_lock_res *res,
+					    struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan_lvb *lvb;
+
+	ocfs2_lock_res_init_once(res);
+	ocfs2_build_lock_name(OCFS2_LOCK_TYPE_ORPHAN_SCAN, 0, 0, res->l_name);
+	ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_ORPHAN_SCAN,
+				   &ocfs2_orphan_scan_lops, osb);
+	lvb = ocfs2_dlm_lvb(&res->l_lksb);
+	lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+}
+
 void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
 			      struct ocfs2_file_private *fp)
 {
@@ -2352,6 +2369,37 @@ void ocfs2_inode_unlock(struct inode *inode,
 	mlog_exit_void();
 }
 
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex)
+{
+	struct ocfs2_lock_res *lockres;
+	struct ocfs2_orphan_scan_lvb *lvb;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+	int status = 0;
+
+	lockres = &osb->osb_orphan_scan.os_lockres;
+	status = ocfs2_cluster_lock(osb, lockres, level, 0, 0); 
+	if (status < 0)
+		return status;
+
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	if (lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
+		*seqno = be32_to_cpu(lvb->lvb_os_seqno); 
+	return status;
+}
+
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, int ex)
+{
+	struct ocfs2_lock_res *lockres;
+	struct ocfs2_orphan_scan_lvb *lvb;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+
+	lockres = &osb->osb_orphan_scan.os_lockres;
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+	lvb->lvb_os_seqno = cpu_to_be32(seqno);
+	ocfs2_cluster_unlock(osb, lockres, level); 
+}
+
 int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex)
 {
@@ -2842,6 +2890,7 @@ local:
 	ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
 	ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
 	ocfs2_nfs_sync_lock_res_init(&osb->osb_nfs_sync_lockres, osb);
+	ocfs2_orphan_scan_lock_res_init(&osb->osb_orphan_scan.os_lockres, osb);
 
 	osb->cconn = conn;
 
@@ -2878,6 +2927,7 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
 	ocfs2_lock_res_free(&osb->osb_super_lockres);
 	ocfs2_lock_res_free(&osb->osb_rename_lockres);
 	ocfs2_lock_res_free(&osb->osb_nfs_sync_lockres);
+	ocfs2_lock_res_free(&osb->osb_orphan_scan.os_lockres);
 
 	ocfs2_cluster_disconnect(osb->cconn, hangup_pending);
 	osb->cconn = NULL;
@@ -3061,6 +3111,7 @@ static void ocfs2_drop_osb_locks(struct ocfs2_super *osb)
 	ocfs2_simple_drop_lockres(osb, &osb->osb_super_lockres);
 	ocfs2_simple_drop_lockres(osb, &osb->osb_rename_lockres);
 	ocfs2_simple_drop_lockres(osb, &osb->osb_nfs_sync_lockres);
+	ocfs2_simple_drop_lockres(osb, &osb->osb_orphan_scan.os_lockres);
 }
 
 int ocfs2_drop_inode_locks(struct inode *inode)
fs/ocfs2/dlmglue.h +11 -0
index e1fd572..31b90d7 100644
@@ -62,6 +62,14 @@ struct ocfs2_qinfo_lvb {
 	__be32	lvb_free_entry;
 };
 
+#define OCFS2_ORPHAN_LVB_VERSION 1
+
+struct ocfs2_orphan_scan_lvb {
+	__u8	lvb_version;
+	__u8	lvb_reserved[3];
+	__be32	lvb_os_seqno;
+};
+
 /* ocfs2_inode_lock_full() 'arg_flags' flags */
 /* don't wait on recovery. */
 #define OCFS2_META_LOCK_RECOVERY	(0x01)
@@ -113,6 +121,9 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex);
 void ocfs2_super_unlock(struct ocfs2_super *osb,
 			int ex);
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex);
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, int ex);
+
 int ocfs2_rename_lock(struct ocfs2_super *osb);
 void ocfs2_rename_unlock(struct ocfs2_super *osb);
 int ocfs2_nfs_sync_lock(struct ocfs2_super *osb, int ex);
fs/ocfs2/journal.c +106 -0
index a20a0f1..dc7cea3 100644
@@ -28,6 +28,8 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/kthread.h>
+#include <linux/time.h>
+#include <linux/random.h>
 
 #define MLOG_MASK_PREFIX ML_JOURNAL
 #include <cluster/masklog.h>
@@ -52,6 +54,8 @@
 
 DEFINE_SPINLOCK(trans_inc_lock);
 
+#define ORPHAN_SCAN_SCHEDULE_TIMEOUT 300000
+
 static int ocfs2_force_read_journal(struct inode *inode);
 static int ocfs2_recover_node(struct ocfs2_super *osb,
 			      int node_num, int slot_num);
@@ -1841,6 +1845,108 @@ bail:
 	return status;
 }
 
+/*
+ * Scan timer should get fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT. Add some
+ * randomness to the timeout to minimize multple nodes firing the timer at the
+ * same time.
+ */
+static inline unsigned long ocfs2_orphan_scan_timeout(void)
+{
+	unsigned long time;
+
+	get_random_bytes(&time, sizeof(time));
+	time = ORPHAN_SCAN_SCHEDULE_TIMEOUT + (time % 5000);
+	return msecs_to_jiffies(time);
+}
+
+/*
+ * ocfs2_queue_orphan_scan calls ocfs2_queue_recovery_completion for
+ * every slot which queues a recovery of slot on ocfs2_wq thread. This is done
+ * to cleanup any orphans that are left over in orphan slots.
+ *
+ * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT seconds
+ * It gets an EX lock on os_lockres and checks sequence number stored in LVB. If
+ * the sequence number is changed it means some node has done the scan. Skip the
+ * scan and tracks the sequence number. If the sequence number didn't change,
+ * means a scan didn't happen, so the node queues a scan and increments the
+ * sequence number in LVB.
+ */ 
+void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+	int status, i;
+	u32 seqno = 0;
+
+	os = &osb->osb_orphan_scan;
+
+	status = ocfs2_orphan_scan_lock(osb, &seqno, DLM_LOCK_EX);
+	if (status < 0) {
+		if (status != -EAGAIN)
+			mlog_errno(status);
+		goto out;
+	}
+
+	if (os->os_seqno != seqno) {
+		os->os_seqno = seqno;
+		goto unlock;
+	}
+
+	for (i = 0; i < osb->max_slots; i++)
+		ocfs2_queue_recovery_completion(osb->journal, i, NULL, NULL,
+						NULL);
+	/*
+	 * We queued a recovery on orphan slots, increment the sequence
+	 * number and update LVB so other node will skip the scan for a while
+	 */
+	seqno++;
+unlock:
+	ocfs2_orphan_scan_unlock(osb, seqno, DLM_LOCK_EX);
+out:
+	return;
+}
+
+/* Worker task that gets fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT millsec */
+void ocfs2_orphan_scan_work(struct work_struct *work)
+{
+	struct ocfs2_orphan_scan *os;
+	struct ocfs2_super *osb;
+
+	os = container_of(work, struct ocfs2_orphan_scan,
+			  os_orphan_scan_work.work);
+	osb = os->os_osb;
+
+	mutex_lock(&os->os_lock);
+	ocfs2_queue_orphan_scan(osb);
+	schedule_delayed_work(&os->os_orphan_scan_work,
+			      ocfs2_orphan_scan_timeout());
+	mutex_unlock(&os->os_lock);
+}
+
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+
+	os = &osb->osb_orphan_scan;
+	mutex_lock(&os->os_lock);
+	cancel_delayed_work(&os->os_orphan_scan_work);
+	mutex_unlock(&os->os_lock);
+}
+
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+
+	os = &osb->osb_orphan_scan;
+	os->os_osb = osb;
+	mutex_init(&os->os_lock);
+
+	INIT_DELAYED_WORK(&os->os_orphan_scan_work,
+			  ocfs2_orphan_scan_work);
+	schedule_delayed_work(&os->os_orphan_scan_work,
+			      ocfs2_orphan_scan_timeout());
+	return 0;
+}
+
 struct ocfs2_orphan_filldir_priv {
 	struct inode		*head;
 	struct ocfs2_super	*osb;
fs/ocfs2/journal.h +4 -0
index 619dd7f..3483202 100644
@@ -144,6 +144,10 @@ static inline void ocfs2_inode_set_new(struct ocfs2_super *osb,
 }
 
 /* Exported only for the journal struct init code in super.c. Do not call. */
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_exit(struct ocfs2_super *osb);
+
 void ocfs2_complete_recovery(struct work_struct *work);
 void ocfs2_wait_for_recovery(struct ocfs2_super *osb);
 
fs/ocfs2/ocfs2.h +10 -0
index 1386281..373fb1c 100644
@@ -151,6 +151,14 @@ struct ocfs2_lock_res {
 #endif
 };
 
+struct ocfs2_orphan_scan {
+	struct mutex 		os_lock;
+	struct ocfs2_super 	*os_osb;
+	struct ocfs2_lock_res 	os_lockres;     /* lock to synchronize scans */
+	struct delayed_work 	os_orphan_scan_work;
+	u32  			os_seqno;       /* incremented on every scan */
+};
+
 struct ocfs2_dlm_debug {
 	struct kref d_refcnt;
 	struct dentry *d_locking_state;
@@ -341,6 +349,8 @@ struct ocfs2_super
 	unsigned int			*osb_orphan_wipes;
 	wait_queue_head_t		osb_wipe_event;
 
+	struct ocfs2_orphan_scan 	osb_orphan_scan; 
+
 	/* used to protect metaecc calculation check of xattr. */
 	spinlock_t osb_xattr_lock;
 
fs/ocfs2/ocfs2_lockid.h +5 -0
index a53ce87..fcdba09 100644
@@ -48,6 +48,7 @@ enum ocfs2_lock_type {
 	OCFS2_LOCK_TYPE_FLOCK,
 	OCFS2_LOCK_TYPE_QINFO,
 	OCFS2_LOCK_TYPE_NFS_SYNC,
+	OCFS2_LOCK_TYPE_ORPHAN_SCAN,
 	OCFS2_NUM_LOCK_TYPES
 };
 
@@ -85,6 +86,9 @@ static inline char ocfs2_lock_type_char(enum ocfs2_lock_type type)
 		case OCFS2_LOCK_TYPE_NFS_SYNC:
 			c = 'Y';
 			break;
+		case OCFS2_LOCK_TYPE_ORPHAN_SCAN:
+			c = 'P';
+			break;
 		default:
 			c = '\0';
 	}
@@ -104,6 +108,7 @@ static char *ocfs2_lock_type_strings[] = {
 	[OCFS2_LOCK_TYPE_OPEN] = "Open",
 	[OCFS2_LOCK_TYPE_FLOCK] = "Flock",
 	[OCFS2_LOCK_TYPE_QINFO] = "Quota",
+	[OCFS2_LOCK_TYPE_ORPHAN_SCAN] = "OrphanScan",
 };
 
 static inline const char *ocfs2_lock_type_string(enum ocfs2_lock_type type)
fs/ocfs2/super.c +9 -0
index 79ff8d9..44ac27e 100644
@@ -1802,6 +1802,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 
 	ocfs2_truncate_log_shutdown(osb);
 
+	ocfs2_orphan_scan_stop(osb);
+
 	/* This will disable recovery and flush any recovery work. */
 	ocfs2_recovery_exit(osb);
 
@@ -1957,6 +1959,13 @@ static int ocfs2_initialize_super(struct super_block *sb,
 		goto bail;
 	}
 
+	status = ocfs2_orphan_scan_init(osb);
+	if (status) {
+		mlog(ML_ERROR, "Unable to initialize delayed orphan scan\n");
+		mlog_errno(status);
+		goto bail;
+	}
+
 	init_waitqueue_head(&osb->checkpoint_event);
 	atomic_set(&osb->needs_checkpoint, 0);
 
#3 lore raw

PATCH 1/2 ocfs2: timer to queue scan of all orphan slots

In reply to Srinivas Eeda

Hi Srini/Joel/Sunil,
I also have some thoughts for it. Wish it isn't too late.

Currently, orphan scan just iterate all the slots and call
ocfs2_queue_recovery_completion, but I don't think it is proper for a node to query another mounted one since that node will query it by
itself. What's more, it will affect reflink greatly.
In my current implementation of reflink, It will work like this:
1. create a inode in orphan dir
2. reflink all the extents.
3. move the inode from orphan dir to the destination.

For efficiency, I just lock orphan dir in step 1 and 3, and release the lock in step 2 in case reflink will take a long time and we don't block other "unlink" process. And in step 1, the created inode looks really like a deleted one so that any crash in step 2 won't prevent it from being deleted by fsck or recovery.

But with your patch, we may have a race in step 2 that your recovery will delete the inode created in step 1. So my suggestion is that your orphan scan just skip the mounted node so it won't affect other nodes' ongoing reflink. As for the node itself, it is very easy to postpone the orphan scan by setting a flag in ocfs2_super when reflink is ongoing(I will do it).

Make sense?

Regards,
Tao

Quoted 386 lines · Srinivas Eeda

When a dentry is unlinked, the unlinking node takes an EX on the dentry lock before moving the dentry to the orphan directory. The other nodes, that all had a PR on the same dentry lock, flag the corresponding inode as MAYBE_ORPHANED during the downconvert. The inode is finally deleted when the last node to iput the inode notices the MAYBE_ORPHANED flag.

A problem arises if a node is forced to free dentry locks because of memory pressure. If this happens, the node will no longer get downconvert notifications for the dentries that have been unlinked on another node. If it also happens that node is actively using the corresponding inode and happens to be the one performing the last iput on that inode, it will fail to delete the inode as it will not have the MAYBE_ORPHANED flag set.

This patch fixes this shortcoming by introducing a periodic scan of the orphan directories to delete such inodes. Care has been taken to distribute the workload across the cluster so that no one node has to perform the task all the time.

---
 fs/ocfs2/dlmglue.c      |   51 ++++++++++++++++++++++
 fs/ocfs2/dlmglue.h      |   11 +++++
 fs/ocfs2/journal.c      |  106 +++++++++++++++++++++++++++++++++++++++++++++++
 fs/ocfs2/journal.h      |    4 ++
 fs/ocfs2/ocfs2.h        |   10 ++++
 fs/ocfs2/ocfs2_lockid.h |    5 ++
 fs/ocfs2/super.c        |    9 ++++
 7 files changed, 196 insertions(+), 0 deletions(-)
fs/ocfs2/dlmglue.c +51 -0
index e15fc7d..0f35b83 100644
@@ -248,6 +248,10 @@ static struct ocfs2_lock_res_ops ocfs2_nfs_sync_lops = {
 	.flags		= 0,
 };
 
+static struct ocfs2_lock_res_ops ocfs2_orphan_scan_lops = {
+	.flags		= LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
+};
+
 static struct ocfs2_lock_res_ops ocfs2_dentry_lops = {
 	.get_osb	= ocfs2_get_dentry_osb,
 	.post_unlock	= ocfs2_dentry_post_unlock,
@@ -637,6 +641,19 @@ static void ocfs2_nfs_sync_lock_res_init(struct ocfs2_lock_res *res,
 				   &ocfs2_nfs_sync_lops, osb);
 }
 
+static void ocfs2_orphan_scan_lock_res_init(struct ocfs2_lock_res *res,
+					    struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan_lvb *lvb;
+
+	ocfs2_lock_res_init_once(res);
+	ocfs2_build_lock_name(OCFS2_LOCK_TYPE_ORPHAN_SCAN, 0, 0, res->l_name);
+	ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_ORPHAN_SCAN,
+				   &ocfs2_orphan_scan_lops, osb);
+	lvb = ocfs2_dlm_lvb(&res->l_lksb);
+	lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+}
+
 void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,
 			      struct ocfs2_file_private *fp)
 {
@@ -2352,6 +2369,37 @@ void ocfs2_inode_unlock(struct inode *inode,
 	mlog_exit_void();
 }
 
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex)
+{
+	struct ocfs2_lock_res *lockres;
+	struct ocfs2_orphan_scan_lvb *lvb;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+	int status = 0;
+
+	lockres = &osb->osb_orphan_scan.os_lockres;
+	status = ocfs2_cluster_lock(osb, lockres, level, 0, 0); 
+	if (status < 0)
+		return status;
+
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	if (lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
+		*seqno = be32_to_cpu(lvb->lvb_os_seqno); 
+	return status;
+}
+
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, int ex)
+{
+	struct ocfs2_lock_res *lockres;
+	struct ocfs2_orphan_scan_lvb *lvb;
+	int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+
+	lockres = &osb->osb_orphan_scan.os_lockres;
+	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+	lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+	lvb->lvb_os_seqno = cpu_to_be32(seqno);
+	ocfs2_cluster_unlock(osb, lockres, level); 
+}
+
 int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex)
 {
@@ -2842,6 +2890,7 @@ local:
 	ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
 	ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
 	ocfs2_nfs_sync_lock_res_init(&osb->osb_nfs_sync_lockres, osb);
+	ocfs2_orphan_scan_lock_res_init(&osb->osb_orphan_scan.os_lockres, osb);
 
 	osb->cconn = conn;
 
@@ -2878,6 +2927,7 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
 	ocfs2_lock_res_free(&osb->osb_super_lockres);
 	ocfs2_lock_res_free(&osb->osb_rename_lockres);
 	ocfs2_lock_res_free(&osb->osb_nfs_sync_lockres);
+	ocfs2_lock_res_free(&osb->osb_orphan_scan.os_lockres);
 
 	ocfs2_cluster_disconnect(osb->cconn, hangup_pending);
 	osb->cconn = NULL;
@@ -3061,6 +3111,7 @@ static void ocfs2_drop_osb_locks(struct ocfs2_super *osb)
 	ocfs2_simple_drop_lockres(osb, &osb->osb_super_lockres);
 	ocfs2_simple_drop_lockres(osb, &osb->osb_rename_lockres);
 	ocfs2_simple_drop_lockres(osb, &osb->osb_nfs_sync_lockres);
+	ocfs2_simple_drop_lockres(osb, &osb->osb_orphan_scan.os_lockres);
 }
 
 int ocfs2_drop_inode_locks(struct inode *inode)
fs/ocfs2/dlmglue.h +11 -0
index e1fd572..31b90d7 100644
@@ -62,6 +62,14 @@ struct ocfs2_qinfo_lvb {
 	__be32	lvb_free_entry;
 };
 
+#define OCFS2_ORPHAN_LVB_VERSION 1
+
+struct ocfs2_orphan_scan_lvb {
+	__u8	lvb_version;
+	__u8	lvb_reserved[3];
+	__be32	lvb_os_seqno;
+};
+
 /* ocfs2_inode_lock_full() 'arg_flags' flags */
 /* don't wait on recovery. */
 #define OCFS2_META_LOCK_RECOVERY	(0x01)
@@ -113,6 +121,9 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
 		     int ex);
 void ocfs2_super_unlock(struct ocfs2_super *osb,
 			int ex);
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex);
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, int ex);
+
 int ocfs2_rename_lock(struct ocfs2_super *osb);
 void ocfs2_rename_unlock(struct ocfs2_super *osb);
 int ocfs2_nfs_sync_lock(struct ocfs2_super *osb, int ex);
fs/ocfs2/journal.c +106 -0
index a20a0f1..dc7cea3 100644
@@ -28,6 +28,8 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/kthread.h>
+#include <linux/time.h>
+#include <linux/random.h>
 
 #define MLOG_MASK_PREFIX ML_JOURNAL
 #include <cluster/masklog.h>
@@ -52,6 +54,8 @@
 
 DEFINE_SPINLOCK(trans_inc_lock);
 
+#define ORPHAN_SCAN_SCHEDULE_TIMEOUT 300000
+
 static int ocfs2_force_read_journal(struct inode *inode);
 static int ocfs2_recover_node(struct ocfs2_super *osb,
 			      int node_num, int slot_num);
@@ -1841,6 +1845,108 @@ bail:
 	return status;
 }
 
+/*
+ * Scan timer should get fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT. Add some
+ * randomness to the timeout to minimize multple nodes firing the timer at the
+ * same time.
+ */
+static inline unsigned long ocfs2_orphan_scan_timeout(void)
+{
+	unsigned long time;
+
+	get_random_bytes(&time, sizeof(time));
+	time = ORPHAN_SCAN_SCHEDULE_TIMEOUT + (time % 5000);
+	return msecs_to_jiffies(time);
+}
+
+/*
+ * ocfs2_queue_orphan_scan calls ocfs2_queue_recovery_completion for
+ * every slot which queues a recovery of slot on ocfs2_wq thread. This is done
+ * to cleanup any orphans that are left over in orphan slots.
+ *
+ * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT seconds
+ * It gets an EX lock on os_lockres and checks sequence number stored in LVB. If
+ * the sequence number is changed it means some node has done the scan. Skip the
+ * scan and tracks the sequence number. If the sequence number didn't change,
+ * means a scan didn't happen, so the node queues a scan and increments the
+ * sequence number in LVB.
+ */ 
+void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+	int status, i;
+	u32 seqno = 0;
+
+	os = &osb->osb_orphan_scan;
+
+	status = ocfs2_orphan_scan_lock(osb, &seqno, DLM_LOCK_EX);
+	if (status < 0) {
+		if (status != -EAGAIN)
+			mlog_errno(status);
+		goto out;
+	}
+
+	if (os->os_seqno != seqno) {
+		os->os_seqno = seqno;
+		goto unlock;
+	}
+
+	for (i = 0; i < osb->max_slots; i++)
+		ocfs2_queue_recovery_completion(osb->journal, i, NULL, NULL,
+						NULL);
+	/*
+	 * We queued a recovery on orphan slots, increment the sequence
+	 * number and update LVB so other node will skip the scan for a while
+	 */
+	seqno++;
+unlock:
+	ocfs2_orphan_scan_unlock(osb, seqno, DLM_LOCK_EX);
+out:
+	return;
+}
+
+/* Worker task that gets fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT millsec */
+void ocfs2_orphan_scan_work(struct work_struct *work)
+{
+	struct ocfs2_orphan_scan *os;
+	struct ocfs2_super *osb;
+
+	os = container_of(work, struct ocfs2_orphan_scan,
+			  os_orphan_scan_work.work);
+	osb = os->os_osb;
+
+	mutex_lock(&os->os_lock);
+	ocfs2_queue_orphan_scan(osb);
+	schedule_delayed_work(&os->os_orphan_scan_work,
+			      ocfs2_orphan_scan_timeout());
+	mutex_unlock(&os->os_lock);
+}
+
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+
+	os = &osb->osb_orphan_scan;
+	mutex_lock(&os->os_lock);
+	cancel_delayed_work(&os->os_orphan_scan_work);
+	mutex_unlock(&os->os_lock);
+}
+
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb)
+{
+	struct ocfs2_orphan_scan *os;
+
+	os = &osb->osb_orphan_scan;
+	os->os_osb = osb;
+	mutex_init(&os->os_lock);
+
+	INIT_DELAYED_WORK(&os->os_orphan_scan_work,
+			  ocfs2_orphan_scan_work);
+	schedule_delayed_work(&os->os_orphan_scan_work,
+			      ocfs2_orphan_scan_timeout());
+	return 0;
+}
+
 struct ocfs2_orphan_filldir_priv {
 	struct inode		*head;
 	struct ocfs2_super	*osb;
fs/ocfs2/journal.h +4 -0
index 619dd7f..3483202 100644
@@ -144,6 +144,10 @@ static inline void ocfs2_inode_set_new(struct ocfs2_super *osb,
 }
 
 /* Exported only for the journal struct init code in super.c. Do not call. */
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_exit(struct ocfs2_super *osb);
+
 void ocfs2_complete_recovery(struct work_struct *work);
 void ocfs2_wait_for_recovery(struct ocfs2_super *osb);
 
fs/ocfs2/ocfs2.h +10 -0
index 1386281..373fb1c 100644
@@ -151,6 +151,14 @@ struct ocfs2_lock_res {
 #endif
 };
 
+struct ocfs2_orphan_scan {
+	struct mutex 		os_lock;
+	struct ocfs2_super 	*os_osb;
+	struct ocfs2_lock_res 	os_lockres;     /* lock to synchronize scans */
+	struct delayed_work 	os_orphan_scan_work;
+	u32  			os_seqno;       /* incremented on every scan */
+};
+
 struct ocfs2_dlm_debug {
 	struct kref d_refcnt;
 	struct dentry *d_locking_state;
@@ -341,6 +349,8 @@ struct ocfs2_super
 	unsigned int			*osb_orphan_wipes;
 	wait_queue_head_t		osb_wipe_event;
 
+	struct ocfs2_orphan_scan 	osb_orphan_scan; 
+
 	/* used to protect metaecc calculation check of xattr. */
 	spinlock_t osb_xattr_lock;
 
fs/ocfs2/ocfs2_lockid.h +5 -0
index a53ce87..fcdba09 100644
@@ -48,6 +48,7 @@ enum ocfs2_lock_type {
 	OCFS2_LOCK_TYPE_FLOCK,
 	OCFS2_LOCK_TYPE_QINFO,
 	OCFS2_LOCK_TYPE_NFS_SYNC,
+	OCFS2_LOCK_TYPE_ORPHAN_SCAN,
 	OCFS2_NUM_LOCK_TYPES
 };
 
@@ -85,6 +86,9 @@ static inline char ocfs2_lock_type_char(enum ocfs2_lock_type type)
 		case OCFS2_LOCK_TYPE_NFS_SYNC:
 			c = 'Y';
 			break;
+		case OCFS2_LOCK_TYPE_ORPHAN_SCAN:
+			c = 'P';
+			break;
 		default:
 			c = '\0';
 	}
@@ -104,6 +108,7 @@ static char *ocfs2_lock_type_strings[] = {
 	[OCFS2_LOCK_TYPE_OPEN] = "Open",
 	[OCFS2_LOCK_TYPE_FLOCK] = "Flock",
 	[OCFS2_LOCK_TYPE_QINFO] = "Quota",
+	[OCFS2_LOCK_TYPE_ORPHAN_SCAN] = "OrphanScan",
 };
 
 static inline const char *ocfs2_lock_type_string(enum ocfs2_lock_type type)
fs/ocfs2/super.c +9 -0
index 79ff8d9..44ac27e 100644
@@ -1802,6 +1802,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 
 	ocfs2_truncate_log_shutdown(osb);
 
+	ocfs2_orphan_scan_stop(osb);
+
 	/* This will disable recovery and flush any recovery work. */
 	ocfs2_recovery_exit(osb);
 
@@ -1957,6 +1959,13 @@ static int ocfs2_initialize_super(struct super_block *sb,
 		goto bail;
 	}
 
+	status = ocfs2_orphan_scan_init(osb);
+	if (status) {
+		mlog(ML_ERROR, "Unable to initialize delayed orphan scan\n");
+		mlog_errno(status);
+		goto bail;
+	}
+
 	init_waitqueue_head(&osb->checkpoint_event);
 	atomic_set(&osb->needs_checkpoint, 0);
 
#4 lore raw

PATCH 1/2 ocfs2: timer to queue scan of all orphan slots

Quoted 18 lines · Tao Ma

Hi Srini/Joel/Sunil,
I also have some thoughts for it. Wish it isn't too late.

Currently, orphan scan just iterate all the slots and call
ocfs2_queue_recovery_completion, but I don't think it is proper for a node to query another mounted one since that node will query it by
itself. What's more, it will affect reflink greatly.
In my current implementation of reflink, It will work like this:
1. create a inode in orphan dir
2. reflink all the extents.
3. move the inode from orphan dir to the destination.

For efficiency, I just lock orphan dir in step 1 and 3, and release the lock in step 2 in case reflink will take a long time and we don't block other "unlink" process. And in step 1, the created inode looks really like a deleted one so that any crash in step 2 won't prevent it from being deleted by fsck or recovery.

Quoted 8 lines

But with your patch, we may have a race in step 2 that your recovery will delete the inode created in step 1. So my suggestion is that your orphan scan just skip the mounted node so it won't affect other nodes' ongoing reflink. As for the node itself, it is very easy to postpone the orphan scan by setting a flag in ocfs2_super when reflink is ongoing(I will do it).

Make sense?

Yes, I can restrict the node to recover it's own and offline slots. I can make the node to recover it's own slot every time the timer fires and offline slots in round robin way(current way)

Quoted 424 lines

Regards,
Tao

Quoted 419 lines · Srinivas Eeda

When a dentry is unlinked, the unlinking node takes an EX on the dentry lock
before moving the dentry to the orphan directory. The other nodes, that all had
a PR on the same dentry lock, flag the corresponding inode as MAYBE_ORPHANED
during the downconvert. The inode is finally deleted when the last node to iput
the inode notices the MAYBE_ORPHANED flag.

A problem arises if a node is forced to free dentry locks because of memory
pressure. If this happens, the node will no longer get downconvert notifications
for the dentries that have been unlinked on another node. If it also happens
that node is actively using the corresponding inode and happens to be the one
performing the last iput on that inode, it will fail to delete the inode as it
will not have the MAYBE_ORPHANED flag set.

This patch fixes this shortcoming by introducing a periodic scan of the orphan
directories to delete such inodes. Care has been taken to distribute the workload across the cluster so that no one node has to perform the task all the
time.

---
 fs/ocfs2/dlmglue.c      |   51 ++++++++++++++++++++++
 fs/ocfs2/dlmglue.h      |   11 +++++
 fs/ocfs2/journal.c      |  106 
+++++++++++++++++++++++++++++++++++++++++++++++
 fs/ocfs2/journal.h      |    4 ++
 fs/ocfs2/ocfs2.h        |   10 ++++
 fs/ocfs2/ocfs2_lockid.h |    5 ++
 fs/ocfs2/super.c        |    9 ++++
 7 files changed, 196 insertions(+), 0 deletions(-)
fs/ocfs2/dlmglue.c +4 -0
index e15fc7d..0f35b83 100644
@@ -248,6 +248,10 @@ static struct ocfs2_lock_res_ops 
 ocfs2_nfs_sync_lops = {
     .flags        = 0,
 };
 
+static struct ocfs2_lock_res_ops ocfs2_orphan_scan_lops = {
+    .flags        = LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
+};
+
 static struct ocfs2_lock_res_ops ocfs2_dentry_lops = {
     .get_osb    = ocfs2_get_dentry_osb,

.post_unlock = ocfs2_dentry_post_unlock,

fs/ocfs2/dlmglue.c +13 -0
@@ -637,6 +641,19 @@ static void ocfs2_nfs_sync_lock_res_init(struct 
 ocfs2_lock_res *res,
                    &ocfs2_nfs_sync_lops, osb);
 }
 
+static void ocfs2_orphan_scan_lock_res_init(struct ocfs2_lock_res *res,
+                        struct ocfs2_super *osb)
+{
+    struct ocfs2_orphan_scan_lvb *lvb;
+
+    ocfs2_lock_res_init_once(res);
+    ocfs2_build_lock_name(OCFS2_LOCK_TYPE_ORPHAN_SCAN, 0, 0, 
 res->l_name);
+    ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_ORPHAN_SCAN,
+                   &ocfs2_orphan_scan_lops, osb);
+    lvb = ocfs2_dlm_lvb(&res->l_lksb);
+    lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+}
+
 void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,

struct ocfs2_file_private *fp)
{

fs/ocfs2/dlmglue.c +31 -0
@@ -2352,6 +2369,37 @@ void ocfs2_inode_unlock(struct inode *inode,
     mlog_exit_void();
 }
 
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex)
+{
+    struct ocfs2_lock_res *lockres;
+    struct ocfs2_orphan_scan_lvb *lvb;
+    int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+    int status = 0;
+
+    lockres = &osb->osb_orphan_scan.os_lockres;
+    status = ocfs2_cluster_lock(osb, lockres, level, 0, 0); +    if 
 (status < 0)
+        return status;
+
+    lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+    if (lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
+        *seqno = be32_to_cpu(lvb->lvb_os_seqno); +    return status;
+}
+
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, 
 int ex)
+{
+    struct ocfs2_lock_res *lockres;
+    struct ocfs2_orphan_scan_lvb *lvb;
+    int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+
+    lockres = &osb->osb_orphan_scan.os_lockres;
+    lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+    lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+    lvb->lvb_os_seqno = cpu_to_be32(seqno);
+    ocfs2_cluster_unlock(osb, lockres, level); +}
+
 int ocfs2_super_lock(struct ocfs2_super *osb,
              int ex)
 {
@@ -2842,6 +2890,7 @@ local:
     ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
     ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
     ocfs2_nfs_sync_lock_res_init(&osb->osb_nfs_sync_lockres, osb);
+    
 ocfs2_orphan_scan_lock_res_init(&osb->osb_orphan_scan.os_lockres, osb);
 
     osb->cconn = conn;
@@ -2878,6 +2927,7 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
     ocfs2_lock_res_free(&osb->osb_super_lockres);
     ocfs2_lock_res_free(&osb->osb_rename_lockres);
     ocfs2_lock_res_free(&osb->osb_nfs_sync_lockres);
+    ocfs2_lock_res_free(&osb->osb_orphan_scan.os_lockres);
 
     ocfs2_cluster_disconnect(osb->cconn, hangup_pending);
     osb->cconn = NULL;
@@ -3061,6 +3111,7 @@ static void ocfs2_drop_osb_locks(struct 
 ocfs2_super *osb)
     ocfs2_simple_drop_lockres(osb, &osb->osb_super_lockres);
     ocfs2_simple_drop_lockres(osb, &osb->osb_rename_lockres);
     ocfs2_simple_drop_lockres(osb, &osb->osb_nfs_sync_lockres);
+    ocfs2_simple_drop_lockres(osb, &osb->osb_orphan_scan.os_lockres);
 }
 

int ocfs2_drop_inode_locks(struct inode *inode)

fs/ocfs2/dlmglue.h +11 -0
index e1fd572..31b90d7 100644
@@ -62,6 +62,14 @@ struct ocfs2_qinfo_lvb {
     __be32    lvb_free_entry;
 };
 
+#define OCFS2_ORPHAN_LVB_VERSION 1
+
+struct ocfs2_orphan_scan_lvb {
+    __u8    lvb_version;
+    __u8    lvb_reserved[3];
+    __be32    lvb_os_seqno;
+};
+
 /* ocfs2_inode_lock_full() 'arg_flags' flags */
 /* don't wait on recovery. */
 #define OCFS2_META_LOCK_RECOVERY    (0x01)
@@ -113,6 +121,9 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
              int ex);
 void ocfs2_super_unlock(struct ocfs2_super *osb,
             int ex);
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int 
 ex);
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, 
 int ex);
+
 int ocfs2_rename_lock(struct ocfs2_super *osb);

void ocfs2_rename_unlock(struct ocfs2_super *osb); int ocfs2_nfs_sync_lock(struct ocfs2_super *osb, int ex);

fs/ocfs2/journal.c +6 -0
index a20a0f1..dc7cea3 100644
@@ -28,6 +28,8 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/kthread.h>
+#include <linux/time.h>
+#include <linux/random.h>
 
 #define MLOG_MASK_PREFIX ML_JOURNAL
 #include <cluster/masklog.h>
@@ -52,6 +54,8 @@
 
 DEFINE_SPINLOCK(trans_inc_lock);
 
+#define ORPHAN_SCAN_SCHEDULE_TIMEOUT 300000
+
 static int ocfs2_force_read_journal(struct inode *inode);
 static int ocfs2_recover_node(struct ocfs2_super *osb,
                   int node_num, int slot_num);
@@ -1841,6 +1845,108 @@ bail:
     return status;
 }
 
+/*
+ * Scan timer should get fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT. 

Add some
+ * randomness to the timeout to minimize multple nodes firing the timer at the
+ * same time.
+ */

fs/ocfs2/journal.c +12 -0
+static inline unsigned long ocfs2_orphan_scan_timeout(void)
+{
+    unsigned long time;
+
+    get_random_bytes(&time, sizeof(time));
+    time = ORPHAN_SCAN_SCHEDULE_TIMEOUT + (time % 5000);
+    return msecs_to_jiffies(time);
+}
+
+/*
+ * ocfs2_queue_orphan_scan calls ocfs2_queue_recovery_completion for
+ * every slot which queues a recovery of slot on ocfs2_wq thread. 

This is done
+ * to cleanup any orphans that are left over in orphan slots.
+ *
+ * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT seconds
+ * It gets an EX lock on os_lockres and checks sequence number stored in LVB. If
+ * the sequence number is changed it means some node has done the scan. Skip the
+ * scan and tracks the sequence number. If the sequence number didn't change,
+ * means a scan didn't happen, so the node queues a scan and increments the
+ * sequence number in LVB.
+ */ +void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)

fs/ocfs2/journal.c +25 -0
+{
+    struct ocfs2_orphan_scan *os;
+    int status, i;
+    u32 seqno = 0;
+
+    os = &osb->osb_orphan_scan;
+
+    status = ocfs2_orphan_scan_lock(osb, &seqno, DLM_LOCK_EX);
+    if (status < 0) {
+        if (status != -EAGAIN)
+            mlog_errno(status);
+        goto out;
+    }
+
+    if (os->os_seqno != seqno) {
+        os->os_seqno = seqno;
+        goto unlock;
+    }
+
+    for (i = 0; i < osb->max_slots; i++)
+        ocfs2_queue_recovery_completion(osb->journal, i, NULL, NULL,
+                        NULL);
+    /*
+     * We queued a recovery on orphan slots, increment the sequence
+     * number and update LVB so other node will skip the scan for a 

while

fs/ocfs2/journal.c +9 -0
+     */
+    seqno++;
+unlock:
+    ocfs2_orphan_scan_unlock(osb, seqno, DLM_LOCK_EX);
+out:
+    return;
+}
+
+/* Worker task that gets fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT 

millsec */

fs/ocfs2/journal.c +41 -0
+void ocfs2_orphan_scan_work(struct work_struct *work)
+{
+    struct ocfs2_orphan_scan *os;
+    struct ocfs2_super *osb;
+
+    os = container_of(work, struct ocfs2_orphan_scan,
+              os_orphan_scan_work.work);
+    osb = os->os_osb;
+
+    mutex_lock(&os->os_lock);
+    ocfs2_queue_orphan_scan(osb);
+    schedule_delayed_work(&os->os_orphan_scan_work,
+                  ocfs2_orphan_scan_timeout());
+    mutex_unlock(&os->os_lock);
+}
+
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb)
+{
+    struct ocfs2_orphan_scan *os;
+
+    os = &osb->osb_orphan_scan;
+    mutex_lock(&os->os_lock);
+    cancel_delayed_work(&os->os_orphan_scan_work);
+    mutex_unlock(&os->os_lock);
+}
+
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb)
+{
+    struct ocfs2_orphan_scan *os;
+
+    os = &osb->osb_orphan_scan;
+    os->os_osb = osb;
+    mutex_init(&os->os_lock);
+
+    INIT_DELAYED_WORK(&os->os_orphan_scan_work,
+              ocfs2_orphan_scan_work);
+    schedule_delayed_work(&os->os_orphan_scan_work,
+                  ocfs2_orphan_scan_timeout());
+    return 0;
+}
+
 struct ocfs2_orphan_filldir_priv {
     struct inode        *head;
     struct ocfs2_super    *osb;
fs/ocfs2/journal.h +0 -0
index 619dd7f..3483202 100644
@@ -144,6 +144,10 @@ static inline void ocfs2_inode_set_new(struct 
 ocfs2_super *osb,
 }
 
 /* Exported only for the journal struct init code in super.c. Do not 

call. */

fs/ocfs2/journal.h +4 -0
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_exit(struct ocfs2_super *osb);
+
 void ocfs2_complete_recovery(struct work_struct *work);
 void ocfs2_wait_for_recovery(struct ocfs2_super *osb);
fs/ocfs2/ocfs2.h +4 -0
index 1386281..373fb1c 100644
@@ -151,6 +151,14 @@ struct ocfs2_lock_res {
 #endif
 };
 
+struct ocfs2_orphan_scan {
+    struct mutex         os_lock;
+    struct ocfs2_super     *os_osb;
+    struct ocfs2_lock_res     os_lockres;     /* lock to synchronize 

scans */

fs/ocfs2/ocfs2.h +5 -0
+    struct delayed_work     os_orphan_scan_work;
+    u32              os_seqno;       /* incremented on every scan */
+};
+
 struct ocfs2_dlm_debug {
     struct kref d_refcnt;
     struct dentry *d_locking_state;
@@ -341,6 +349,8 @@ struct ocfs2_super
     unsigned int            *osb_orphan_wipes;
     wait_queue_head_t        osb_wipe_event;
 
+    struct ocfs2_orphan_scan     osb_orphan_scan; +
     /* used to protect metaecc calculation check of xattr. */
     spinlock_t osb_xattr_lock;
 
fs/ocfs2/ocfs2_lockid.h +4 -0
index a53ce87..fcdba09 100644
@@ -48,6 +48,7 @@ enum ocfs2_lock_type {
     OCFS2_LOCK_TYPE_FLOCK,
     OCFS2_LOCK_TYPE_QINFO,
     OCFS2_LOCK_TYPE_NFS_SYNC,
+    OCFS2_LOCK_TYPE_ORPHAN_SCAN,
     OCFS2_NUM_LOCK_TYPES
 };
 
@@ -85,6 +86,9 @@ static inline char ocfs2_lock_type_char(enum 
 ocfs2_lock_type type)
         case OCFS2_LOCK_TYPE_NFS_SYNC:
             c = 'Y';
             break;
+        case OCFS2_LOCK_TYPE_ORPHAN_SCAN:
+            c = 'P';
+            break;
         default:
             c = '\0';
     }
fs/ocfs2/ocfs2_lockid.h +1 -0
@@ -104,6 +108,7 @@ static char *ocfs2_lock_type_strings[] = {
     [OCFS2_LOCK_TYPE_OPEN] = "Open",
     [OCFS2_LOCK_TYPE_FLOCK] = "Flock",
     [OCFS2_LOCK_TYPE_QINFO] = "Quota",
+    [OCFS2_LOCK_TYPE_ORPHAN_SCAN] = "OrphanScan",
 };
 
 static inline const char *ocfs2_lock_type_string(enum 

ocfs2_lock_type type)

fs/ocfs2/super.c +9 -0
index 79ff8d9..44ac27e 100644
@@ -1802,6 +1802,8 @@ static void ocfs2_dismount_volume(struct 
 super_block *sb, int mnt_err)
 
     ocfs2_truncate_log_shutdown(osb);
 
+    ocfs2_orphan_scan_stop(osb);
+
     /* This will disable recovery and flush any recovery work. */
     ocfs2_recovery_exit(osb);
@@ -1957,6 +1959,13 @@ static int ocfs2_initialize_super(struct 
 super_block *sb,
         goto bail;
     }
 
+    status = ocfs2_orphan_scan_init(osb);
+    if (status) {
+        mlog(ML_ERROR, "Unable to initialize delayed orphan scan\n");
+        mlog_errno(status);
+        goto bail;
+    }
+
     init_waitqueue_head(&osb->checkpoint_event);
     atomic_set(&osb->needs_checkpoint, 0);
#5 lore raw

PATCH 1/2 ocfs2: timer to queue scan of all orphan slots

Quoted 26 lines · Srinivas Eeda
Quoted 25 lines · Tao Ma

Hi Srini/Joel/Sunil,
I also have some thoughts for it. Wish it isn't too late.

Currently, orphan scan just iterate all the slots and call
ocfs2_queue_recovery_completion, but I don't think it is proper for a node to query another mounted one since that node will query it by
itself. What's more, it will affect reflink greatly.
In my current implementation of reflink, It will work like this:
1. create a inode in orphan dir
2. reflink all the extents.
3. move the inode from orphan dir to the destination.

For efficiency, I just lock orphan dir in step 1 and 3, and release the lock in step 2 in case reflink will take a long time and we don't block other "unlink" process. And in step 1, the created inode looks really like a deleted one so that any crash in step 2 won't prevent it from being deleted by fsck or recovery.

But with your patch, we may have a race in step 2 that your recovery will delete the inode created in step 1. So my suggestion is that your orphan scan just skip the mounted node so it won't affect other nodes' ongoing reflink. As for the node itself, it is very easy to postpone the orphan scan by setting a flag in ocfs2_super when reflink is ongoing(I will do it).

Quoted 4 lines
Quoted 1 line

Make sense?

Yes, I can restrict the node to recover it's own and offline slots. I can make the node to recover it's own slot every time the timer fires and offline slots in round robin way(current way)

yeah, that would be cool.

Regards,
Tao

Quoted 424 lines
Quoted 424 lines

Regards,
Tao

Quoted 419 lines · Srinivas Eeda

When a dentry is unlinked, the unlinking node takes an EX on the dentry lock
before moving the dentry to the orphan directory. The other nodes, that all had
a PR on the same dentry lock, flag the corresponding inode as MAYBE_ORPHANED
during the downconvert. The inode is finally deleted when the last node to iput
the inode notices the MAYBE_ORPHANED flag.

A problem arises if a node is forced to free dentry locks because of memory
pressure. If this happens, the node will no longer get downconvert notifications
for the dentries that have been unlinked on another node. If it also happens
that node is actively using the corresponding inode and happens to be the one
performing the last iput on that inode, it will fail to delete the inode as it
will not have the MAYBE_ORPHANED flag set.

This patch fixes this shortcoming by introducing a periodic scan of the orphan
directories to delete such inodes. Care has been taken to distribute the workload across the cluster so that no one node has to perform the task all the
time.

---
 fs/ocfs2/dlmglue.c      |   51 ++++++++++++++++++++++
 fs/ocfs2/dlmglue.h      |   11 +++++
 fs/ocfs2/journal.c      |  106 
+++++++++++++++++++++++++++++++++++++++++++++++
 fs/ocfs2/journal.h      |    4 ++
 fs/ocfs2/ocfs2.h        |   10 ++++
 fs/ocfs2/ocfs2_lockid.h |    5 ++
 fs/ocfs2/super.c        |    9 ++++
 7 files changed, 196 insertions(+), 0 deletions(-)
fs/ocfs2/dlmglue.c +4 -0
index e15fc7d..0f35b83 100644
@@ -248,6 +248,10 @@ static struct ocfs2_lock_res_ops 
 ocfs2_nfs_sync_lops = {
     .flags        = 0,
 };
 
+static struct ocfs2_lock_res_ops ocfs2_orphan_scan_lops = {
+    .flags        = LOCK_TYPE_REQUIRES_REFRESH|LOCK_TYPE_USES_LVB,
+};
+
 static struct ocfs2_lock_res_ops ocfs2_dentry_lops = {
     .get_osb    = ocfs2_get_dentry_osb,

.post_unlock = ocfs2_dentry_post_unlock,

fs/ocfs2/dlmglue.c +13 -0
@@ -637,6 +641,19 @@ static void ocfs2_nfs_sync_lock_res_init(struct 
 ocfs2_lock_res *res,
                    &ocfs2_nfs_sync_lops, osb);
 }
 
+static void ocfs2_orphan_scan_lock_res_init(struct ocfs2_lock_res *res,
+                        struct ocfs2_super *osb)
+{
+    struct ocfs2_orphan_scan_lvb *lvb;
+
+    ocfs2_lock_res_init_once(res);
+    ocfs2_build_lock_name(OCFS2_LOCK_TYPE_ORPHAN_SCAN, 0, 0, 
 res->l_name);
+    ocfs2_lock_res_init_common(osb, res, OCFS2_LOCK_TYPE_ORPHAN_SCAN,
+                   &ocfs2_orphan_scan_lops, osb);
+    lvb = ocfs2_dlm_lvb(&res->l_lksb);
+    lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+}
+
 void ocfs2_file_lock_res_init(struct ocfs2_lock_res *lockres,

struct ocfs2_file_private *fp)
{

fs/ocfs2/dlmglue.c +31 -0
@@ -2352,6 +2369,37 @@ void ocfs2_inode_unlock(struct inode *inode,
     mlog_exit_void();
 }
 
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex)
+{
+    struct ocfs2_lock_res *lockres;
+    struct ocfs2_orphan_scan_lvb *lvb;
+    int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+    int status = 0;
+
+    lockres = &osb->osb_orphan_scan.os_lockres;
+    status = ocfs2_cluster_lock(osb, lockres, level, 0, 0); +    if 
 (status < 0)
+        return status;
+
+    lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+    if (lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
+        *seqno = be32_to_cpu(lvb->lvb_os_seqno); +    return status;
+}
+
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, 
 int ex)
+{
+    struct ocfs2_lock_res *lockres;
+    struct ocfs2_orphan_scan_lvb *lvb;
+    int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR;
+
+    lockres = &osb->osb_orphan_scan.os_lockres;
+    lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
+    lvb->lvb_version = OCFS2_ORPHAN_LVB_VERSION;
+    lvb->lvb_os_seqno = cpu_to_be32(seqno);
+    ocfs2_cluster_unlock(osb, lockres, level); +}
+
 int ocfs2_super_lock(struct ocfs2_super *osb,
              int ex)
 {
@@ -2842,6 +2890,7 @@ local:
     ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb);
     ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb);
     ocfs2_nfs_sync_lock_res_init(&osb->osb_nfs_sync_lockres, osb);
+    
 ocfs2_orphan_scan_lock_res_init(&osb->osb_orphan_scan.os_lockres, osb);
 
     osb->cconn = conn;
@@ -2878,6 +2927,7 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb,
     ocfs2_lock_res_free(&osb->osb_super_lockres);
     ocfs2_lock_res_free(&osb->osb_rename_lockres);
     ocfs2_lock_res_free(&osb->osb_nfs_sync_lockres);
+    ocfs2_lock_res_free(&osb->osb_orphan_scan.os_lockres);
 
     ocfs2_cluster_disconnect(osb->cconn, hangup_pending);
     osb->cconn = NULL;
@@ -3061,6 +3111,7 @@ static void ocfs2_drop_osb_locks(struct 
 ocfs2_super *osb)
     ocfs2_simple_drop_lockres(osb, &osb->osb_super_lockres);
     ocfs2_simple_drop_lockres(osb, &osb->osb_rename_lockres);
     ocfs2_simple_drop_lockres(osb, &osb->osb_nfs_sync_lockres);
+    ocfs2_simple_drop_lockres(osb, &osb->osb_orphan_scan.os_lockres);
 }
 

int ocfs2_drop_inode_locks(struct inode *inode)

fs/ocfs2/dlmglue.h +11 -0
index e1fd572..31b90d7 100644
@@ -62,6 +62,14 @@ struct ocfs2_qinfo_lvb {
     __be32    lvb_free_entry;
 };
 
+#define OCFS2_ORPHAN_LVB_VERSION 1
+
+struct ocfs2_orphan_scan_lvb {
+    __u8    lvb_version;
+    __u8    lvb_reserved[3];
+    __be32    lvb_os_seqno;
+};
+
 /* ocfs2_inode_lock_full() 'arg_flags' flags */
 /* don't wait on recovery. */
 #define OCFS2_META_LOCK_RECOVERY    (0x01)
@@ -113,6 +121,9 @@ int ocfs2_super_lock(struct ocfs2_super *osb,
              int ex);
 void ocfs2_super_unlock(struct ocfs2_super *osb,
             int ex);
+int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int 
 ex);
+void ocfs2_orphan_scan_unlock(struct ocfs2_super *osb, u32 seqno, 
 int ex);
+
 int ocfs2_rename_lock(struct ocfs2_super *osb);

void ocfs2_rename_unlock(struct ocfs2_super *osb); int ocfs2_nfs_sync_lock(struct ocfs2_super *osb, int ex);

fs/ocfs2/journal.c +6 -0
index a20a0f1..dc7cea3 100644
@@ -28,6 +28,8 @@
 #include <linux/slab.h>
 #include <linux/highmem.h>
 #include <linux/kthread.h>
+#include <linux/time.h>
+#include <linux/random.h>
 
 #define MLOG_MASK_PREFIX ML_JOURNAL
 #include <cluster/masklog.h>
@@ -52,6 +54,8 @@
 
 DEFINE_SPINLOCK(trans_inc_lock);
 
+#define ORPHAN_SCAN_SCHEDULE_TIMEOUT 300000
+
 static int ocfs2_force_read_journal(struct inode *inode);
 static int ocfs2_recover_node(struct ocfs2_super *osb,
                   int node_num, int slot_num);
@@ -1841,6 +1845,108 @@ bail:
     return status;
 }
 
+/*
+ * Scan timer should get fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT. 

Add some
+ * randomness to the timeout to minimize multple nodes firing the timer at the
+ * same time.
+ */

fs/ocfs2/journal.c +12 -0
+static inline unsigned long ocfs2_orphan_scan_timeout(void)
+{
+    unsigned long time;
+
+    get_random_bytes(&time, sizeof(time));
+    time = ORPHAN_SCAN_SCHEDULE_TIMEOUT + (time % 5000);
+    return msecs_to_jiffies(time);
+}
+
+/*
+ * ocfs2_queue_orphan_scan calls ocfs2_queue_recovery_completion for
+ * every slot which queues a recovery of slot on ocfs2_wq thread. 

This is done
+ * to cleanup any orphans that are left over in orphan slots.
+ *
+ * ocfs2_queue_orphan_scan gets called every ORPHAN_SCAN_SCHEDULE_TIMEOUT seconds
+ * It gets an EX lock on os_lockres and checks sequence number stored in LVB. If
+ * the sequence number is changed it means some node has done the scan. Skip the
+ * scan and tracks the sequence number. If the sequence number didn't change,
+ * means a scan didn't happen, so the node queues a scan and increments the
+ * sequence number in LVB.
+ */ +void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)

fs/ocfs2/journal.c +25 -0
+{
+    struct ocfs2_orphan_scan *os;
+    int status, i;
+    u32 seqno = 0;
+
+    os = &osb->osb_orphan_scan;
+
+    status = ocfs2_orphan_scan_lock(osb, &seqno, DLM_LOCK_EX);
+    if (status < 0) {
+        if (status != -EAGAIN)
+            mlog_errno(status);
+        goto out;
+    }
+
+    if (os->os_seqno != seqno) {
+        os->os_seqno = seqno;
+        goto unlock;
+    }
+
+    for (i = 0; i < osb->max_slots; i++)
+        ocfs2_queue_recovery_completion(osb->journal, i, NULL, NULL,
+                        NULL);
+    /*
+     * We queued a recovery on orphan slots, increment the sequence
+     * number and update LVB so other node will skip the scan for a 

while

fs/ocfs2/journal.c +9 -0
+     */
+    seqno++;
+unlock:
+    ocfs2_orphan_scan_unlock(osb, seqno, DLM_LOCK_EX);
+out:
+    return;
+}
+
+/* Worker task that gets fired every ORPHAN_SCAN_SCHEDULE_TIMEOUT 

millsec */

fs/ocfs2/journal.c +41 -0
+void ocfs2_orphan_scan_work(struct work_struct *work)
+{
+    struct ocfs2_orphan_scan *os;
+    struct ocfs2_super *osb;
+
+    os = container_of(work, struct ocfs2_orphan_scan,
+              os_orphan_scan_work.work);
+    osb = os->os_osb;
+
+    mutex_lock(&os->os_lock);
+    ocfs2_queue_orphan_scan(osb);
+    schedule_delayed_work(&os->os_orphan_scan_work,
+                  ocfs2_orphan_scan_timeout());
+    mutex_unlock(&os->os_lock);
+}
+
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb)
+{
+    struct ocfs2_orphan_scan *os;
+
+    os = &osb->osb_orphan_scan;
+    mutex_lock(&os->os_lock);
+    cancel_delayed_work(&os->os_orphan_scan_work);
+    mutex_unlock(&os->os_lock);
+}
+
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb)
+{
+    struct ocfs2_orphan_scan *os;
+
+    os = &osb->osb_orphan_scan;
+    os->os_osb = osb;
+    mutex_init(&os->os_lock);
+
+    INIT_DELAYED_WORK(&os->os_orphan_scan_work,
+              ocfs2_orphan_scan_work);
+    schedule_delayed_work(&os->os_orphan_scan_work,
+                  ocfs2_orphan_scan_timeout());
+    return 0;
+}
+
 struct ocfs2_orphan_filldir_priv {
     struct inode        *head;
     struct ocfs2_super    *osb;
fs/ocfs2/journal.h +0 -0
index 619dd7f..3483202 100644
@@ -144,6 +144,10 @@ static inline void ocfs2_inode_set_new(struct 
 ocfs2_super *osb,
 }
 
 /* Exported only for the journal struct init code in super.c. Do not 

call. */

fs/ocfs2/journal.h +4 -0
+int ocfs2_orphan_scan_init(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_stop(struct ocfs2_super *osb);
+void ocfs2_orphan_scan_exit(struct ocfs2_super *osb);
+
 void ocfs2_complete_recovery(struct work_struct *work);
 void ocfs2_wait_for_recovery(struct ocfs2_super *osb);
fs/ocfs2/ocfs2.h +4 -0
index 1386281..373fb1c 100644
@@ -151,6 +151,14 @@ struct ocfs2_lock_res {
 #endif
 };
 
+struct ocfs2_orphan_scan {
+    struct mutex         os_lock;
+    struct ocfs2_super     *os_osb;
+    struct ocfs2_lock_res     os_lockres;     /* lock to synchronize 

scans */

fs/ocfs2/ocfs2.h +5 -0
+    struct delayed_work     os_orphan_scan_work;
+    u32              os_seqno;       /* incremented on every scan */
+};
+
 struct ocfs2_dlm_debug {
     struct kref d_refcnt;
     struct dentry *d_locking_state;
@@ -341,6 +349,8 @@ struct ocfs2_super
     unsigned int            *osb_orphan_wipes;
     wait_queue_head_t        osb_wipe_event;
 
+    struct ocfs2_orphan_scan     osb_orphan_scan; +
     /* used to protect metaecc calculation check of xattr. */
     spinlock_t osb_xattr_lock;
 
fs/ocfs2/ocfs2_lockid.h +4 -0
index a53ce87..fcdba09 100644
@@ -48,6 +48,7 @@ enum ocfs2_lock_type {
     OCFS2_LOCK_TYPE_FLOCK,
     OCFS2_LOCK_TYPE_QINFO,
     OCFS2_LOCK_TYPE_NFS_SYNC,
+    OCFS2_LOCK_TYPE_ORPHAN_SCAN,
     OCFS2_NUM_LOCK_TYPES
 };
 
@@ -85,6 +86,9 @@ static inline char ocfs2_lock_type_char(enum 
 ocfs2_lock_type type)
         case OCFS2_LOCK_TYPE_NFS_SYNC:
             c = 'Y';
             break;
+        case OCFS2_LOCK_TYPE_ORPHAN_SCAN:
+            c = 'P';
+            break;
         default:
             c = '\0';
     }
fs/ocfs2/ocfs2_lockid.h +1 -0
@@ -104,6 +108,7 @@ static char *ocfs2_lock_type_strings[] = {
     [OCFS2_LOCK_TYPE_OPEN] = "Open",
     [OCFS2_LOCK_TYPE_FLOCK] = "Flock",
     [OCFS2_LOCK_TYPE_QINFO] = "Quota",
+    [OCFS2_LOCK_TYPE_ORPHAN_SCAN] = "OrphanScan",
 };
 
 static inline const char *ocfs2_lock_type_string(enum 

ocfs2_lock_type type)

fs/ocfs2/super.c +9 -0
index 79ff8d9..44ac27e 100644
@@ -1802,6 +1802,8 @@ static void ocfs2_dismount_volume(struct 
 super_block *sb, int mnt_err)
 
     ocfs2_truncate_log_shutdown(osb);
 
+    ocfs2_orphan_scan_stop(osb);
+
     /* This will disable recovery and flush any recovery work. */
     ocfs2_recovery_exit(osb);
@@ -1957,6 +1959,13 @@ static int ocfs2_initialize_super(struct 
 super_block *sb,
         goto bail;
     }
 
+    status = ocfs2_orphan_scan_init(osb);
+    if (status) {
+        mlog(ML_ERROR, "Unable to initialize delayed orphan scan\n");
+        mlog_errno(status);
+        goto bail;
+    }
+
     init_waitqueue_head(&osb->checkpoint_event);
     atomic_set(&osb->needs_checkpoint, 0);
#6 lore raw

PATCH 1/2 ocfs2: timer to queue scan of all orphan slots

In reply to Tao Ma
Quoted 1 line · Tao Ma

I also have some thoughts for it. Wish it isn't too late.

Well, if we come up with changes it will affect what I push, but
that's OK.

Quoted 4 lines

Currently, orphan scan just iterate all the slots and call
ocfs2_queue_recovery_completion, but I don't think it is proper for a node to query another mounted one since that node will query it by itself.

Node 1 has an inode it was using. The dentry went away due to
memory pressure. Node 1 closes the inode, but it's on the free list. The node has the open lock.
Node 2 unlinks the inode. It grabs the dentry lock to notify
others, but node 1 has no dentry and doesn't get the message. It trylocks the open lock, sees that another node has a PR, and does nothing.
Later node 2 runs its orphan dir. It igets the inode, trylocks
the open lock, sees the PR still, and does nothing.
Basically, we have to trigger an orphan iput on node 1. The
only way for this to happen is if node 1 runs node 2's orphan dir. This patch exists because that wasn't happening.

Quoted 12 lines

What's more, it will affect reflink greatly.
In my current implementation of reflink, It will work like this:
1. create a inode in orphan dir
2. reflink all the extents.
3. move the inode from orphan dir to the destination.

For efficiency, I just lock orphan dir in step 1 and 3, and release the lock in step 2 in case reflink will take a long time and we don't block other "unlink" process. And in step 1, the created inode looks really like a deleted one so that any crash in step 2 won't prevent it from being deleted by fsck or recovery.

Quoted 6 lines

But with your patch, we may have a race in step 2 that your recovery will delete the inode created in step 1. So my suggestion is that your orphan scan just skip the mounted node so it won't affect other nodes' ongoing reflink. As for the node itself, it is very easy to postpone the orphan scan by setting a flag in ocfs2_super when reflink is ongoing(I will do it).

You should have an in-core inode, right? That holds the open
lock, preventing the others from deleting it. If you crash, then your open lock goes away, and it can be recovered.
More importantly, your orphan dir can be run on regular recovery
async as well. It has to work in all cases.

Joel

Signature · 10 lines
"I'm drifting and drifting
 Just like a ship out on the sea.
 Cause I ain't got nobody, baby,
 In this world to care for me."

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127
#7 lore raw

PATCH 1/2 ocfs2: timer to queue scan of all orphan slots

Hi Joel,
This reply may be really too late. :)

Quoted 11 lines · Joel Becker
Quoted 1 line · Tao Ma

I also have some thoughts for it. Wish it isn't too late.

Well, if we come up with changes it will affect what I push, but
that's OK.

Quoted 4 lines

Currently, orphan scan just iterate all the slots and call
ocfs2_queue_recovery_completion, but I don't think it is proper for a node to query another mounted one since that node will query it by itself.

Quoted 7 lines

Node 1 has an inode it was using. The dentry went away due to
memory pressure. Node 1 closes the inode, but it's on the free list. The node has the open lock.
Node 2 unlinks the inode. It grabs the dentry lock to notify
others, but node 1 has no dentry and doesn't get the message. It trylocks the open lock, sees that another node has a PR, and does nothing.

I just went through the codes of orphan delete, and I think in this case, we should have already released the open lock in node 1? When dentry in node 1 went away, it iput. And when node 1 close the inode, it iputs and open_lock is unlocked already. So node 2 should be OK to delete the file.

I guess the only case orphan scan help is that dentry in node 1 went away while the file is opened and at that time node 2 unlink the file. Am I wrong?

Quoted 5 lines

Later node 2 runs its orphan dir. It igets the inode, trylocks
the open lock, sees the PR still, and does nothing.
Basically, we have to trigger an orphan iput on node 1. The
only way for this to happen is if node 1 runs node 2's orphan dir. This patch exists because that wasn't happening.

If the above case I described is right, orphan scan would work after node 1 close the inode. node 2 will scan its slot, and then try iget->iput->try_open_lock->delete_inode, the file will be deleted finally. So we won't trigger an iput in node1.

Quoted 20 lines
Quoted 13 lines

What's more, it will affect reflink greatly.
In my current implementation of reflink, It will work like this:
1. create a inode in orphan dir
2. reflink all the extents.
3. move the inode from orphan dir to the destination.

For efficiency, I just lock orphan dir in step 1 and 3, and release the lock in step 2 in case reflink will take a long time and we don't block other "unlink" process. And in step 1, the created inode looks really like a deleted one so that any crash in step 2 won't prevent it from being deleted by fsck or recovery.

Quoted 6 lines

But with your patch, we may have a race in step 2 that your recovery will delete the inode created in step 1. So my suggestion is that your orphan scan just skip the mounted node so it won't affect other nodes' ongoing reflink. As for the node itself, it is very easy to postpone the orphan scan by setting a flag in ocfs2_super when reflink is ongoing(I will do it).

Quoted 5 lines

You should have an in-core inode, right? That holds the open
lock, preventing the others from deleting it. If you crash, then your open lock goes away, and it can be recovered.
More importantly, your orphan dir can be run on regular recovery
async as well. It has to work in all cases.

yes, I have already added open_lock. So orphan scan won't affect reflink actually. I just want to clarify the scenario orphan scan really works. ;)

Regards,
Tao

#8 lore raw

PATCH 1/2 ocfs2: timer to queue scan of all orphan slots

Quoted 28 lines · Tao Ma

Hi Joel,
This reply may be really too late. :)

Quoted 11 lines · Joel Becker
Quoted 1 line · Tao Ma

I also have some thoughts for it. Wish it isn't too late.

Well, if we come up with changes it will affect what I push, but
that's OK.

Quoted 4 lines

Currently, orphan scan just iterate all the slots and call
ocfs2_queue_recovery_completion, but I don't think it is proper for a node to query another mounted one since that node will query it by itself.

Quoted 7 lines

Node 1 has an inode it was using. The dentry went away due to
memory pressure. Node 1 closes the inode, but it's on the free list. The node has the open lock.
Node 2 unlinks the inode. It grabs the dentry lock to notify
others, but node 1 has no dentry and doesn't get the message. It trylocks the open lock, sees that another node has a PR, and does nothing.

I just went through the codes of orphan delete, and I think in this case, we should have already released the open lock in node 1? When dentry in node 1 went away, it iput. And when node 1 close the inode, it iputs and open_lock is unlocked already. So node 2 should be OK to delete the file.

Quoted 3 lines

I guess the only case orphan scan help is that dentry in node 1 went away while the file is opened and at that time node 2 unlink the file. Am I wrong?

correct, but the file may not be opened. inode is node 1's cache.

Quoted 9 lines
Quoted 5 lines

Later node 2 runs its orphan dir. It igets the inode, trylocks
the open lock, sees the PR still, and does nothing.
Basically, we have to trigger an orphan iput on node 1. The
only way for this to happen is if node 1 runs node 2's orphan dir. This patch exists because that wasn't happening.

If the above case I described is right, orphan scan would work after node 1 close the inode. node 2 will scan its slot, and then try iget->iput->try_open_lock->delete_inode, the file will be deleted finally. So we won't trigger an iput in node1.

yes, the only problem is inode could be in node 1's cache for a very long time. But yes, once after node 1 flushes the inode and node 2 scans the slot it will be able to delete the file. In a multiple nodes cases, inode could be in multiple nodes cache.

Quoted 31 lines
Quoted 20 lines
Quoted 13 lines

What's more, it will affect reflink greatly.
In my current implementation of reflink, It will work like this:
1. create a inode in orphan dir
2. reflink all the extents.
3. move the inode from orphan dir to the destination.

For efficiency, I just lock orphan dir in step 1 and 3, and release the lock in step 2 in case reflink will take a long time and we don't block other "unlink" process. And in step 1, the created inode looks really like a deleted one so that any crash in step 2 won't prevent it from being deleted by fsck or recovery.

Quoted 6 lines

But with your patch, we may have a race in step 2 that your recovery will delete the inode created in step 1. So my suggestion is that your orphan scan just skip the mounted node so it won't affect other nodes' ongoing reflink. As for the node itself, it is very easy to postpone the orphan scan by setting a flag in ocfs2_super when reflink is ongoing(I will do it).

Quoted 5 lines

You should have an in-core inode, right? That holds the open
lock, preventing the others from deleting it. If you crash, then your open lock goes away, and it can be recovered.
More importantly, your orphan dir can be run on regular recovery
async as well. It has to work in all cases.

yes, I have already added open_lock. So orphan scan won't affect reflink actually. I just want to clarify the scenario orphan scan really works. ;)

Regards,
Tao

#9 lore raw

PATCH 2/2 ocfs2 patch to track delayed orphan scan timer statistics

In reply to Srinivas Eeda

Patch to track delayed orphan scan timer statistics.

Modifies ocfs2_osb_dump to print the following:
Orphan Scan=> Local: 10 Global: 21 Last Scan: 67 seconds ago

---
 fs/ocfs2/journal.c |    4 ++++
 fs/ocfs2/ocfs2.h   |    4 +++-
 fs/ocfs2/super.c   |    8 ++++++++
 3 files changed, 15 insertions(+), 1 deletions(-)
fs/ocfs2/journal.c +4 -0
index dc7cea3..4d2e28c 100644
@@ -1899,6 +1899,8 @@ void ocfs2_queue_orphan_scan(struct ocfs2_super *osb)
 	 * number and update LVB so other node will skip the scan for a while
 	 */
 	seqno++;
+	os->os_count++;
+	os->os_scantime = CURRENT_TIME;
 unlock:
 	ocfs2_orphan_scan_unlock(osb, seqno, DLM_LOCK_EX);
 out:
@@ -1938,6 +1940,8 @@ int ocfs2_orphan_scan_init(struct ocfs2_super *osb)
 
 	os = &osb->osb_orphan_scan;
 	os->os_osb = osb;
+	os->os_count = 0;
+	os->os_scantime = CURRENT_TIME;
 	mutex_init(&os->os_lock);
 
 	INIT_DELAYED_WORK(&os->os_orphan_scan_work,
fs/ocfs2/ocfs2.h +3 -1
index 373fb1c..3355b1a 100644
@@ -156,7 +156,9 @@ struct ocfs2_orphan_scan {
 	struct ocfs2_super 	*os_osb;
 	struct ocfs2_lock_res 	os_lockres;     /* lock to synchronize scans */
 	struct delayed_work 	os_orphan_scan_work;
-	u32  			os_seqno;       /* incremented on every scan */
+	struct timespec		os_scantime;  /* time this node ran the scan */
+	u32			os_count;      /* tracks node specific scans */
+	u32  			os_seqno;       /* tracks cluster wide scans */
 };
 
 struct ocfs2_dlm_debug {
fs/ocfs2/super.c +8 -0
index 44ac27e..d05f3ca 100644
@@ -208,6 +208,7 @@ static int ocfs2_osb_dump(struct ocfs2_super *osb, char *buf, int len)
 	int i;
 	struct ocfs2_cluster_connection *cconn = osb->cconn;
 	struct ocfs2_recovery_map *rm = osb->recovery_map;
+	struct ocfs2_orphan_scan *os;
 
 	out += snprintf(buf + out, len - out,
 			"%10s => Id: %-s  Uuid: %-s  Gen: 0x%X  Label: %-s\n",
@@ -309,6 +310,13 @@ static int ocfs2_osb_dump(struct ocfs2_super *osb, char *buf, int len)
 				i, osb->slot_recovery_generations[i]);
 	}
 
+	os = &osb->osb_orphan_scan;
+	out += snprintf(buf + out, len - out, "Orphan Scan=> ");
+	out += snprintf(buf + out, len - out, "Local: %u  Global: %u ",
+			os->os_count, os->os_seqno);
+	out += snprintf(buf + out, len - out, " Last Scan: %lu seconds ago\n",
+			(get_seconds() - os->os_scantime.tv_sec));
+
 	return out;
 }
 
Signature · 1 line
1.5.6.5
#10 lore raw
In reply to Srinivas Eeda
Quoted 1 line · Srinivas Eeda

Resending after implementing review comments.

These patches are now in the merge-window branch of ocfs2.git.
They also live on the orphan-scan topic branch.

Joel

Signature · 9 lines
"I think it would be a good idea."  
        - Mahatma Ghandi, when asked what he thought of Western
          civilization

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127