light: emergency stop logging on filesystem full

mars_trans_logger.c

@@ -625,7 +625,7 @@ int trans_logger_ref_get(struct trans_logger_output *output, struct mref_object
 		mref->ref_len = REGION_SIZE - base_offset;
 	}
 
-	if (mref->ref_may_write == READ) {
+	if (mref->ref_may_write == READ || unlikely(brick->cease_logging)) {
 		return _read_ref_get(output, mref_a);
 	}
 
@@ -824,7 +824,7 @@ void trans_logger_ref_io(struct trans_logger_output *output, struct mref_object
 	}
 
 	// only READ is allowed on non-shadow buffers
-	if (unlikely(mref->ref_rw != READ)) {
+	if (unlikely(mref->ref_rw != READ && !brick->cease_logging)) {
 		MARS_FAT("bad operation %d on non-shadow\n", mref->ref_rw);
 	}
 

mars_trans_logger.h

@@ -120,6 +120,7 @@ struct trans_logger_brick {
 	bool replay_mode;   // mode of operation
 	bool continuous_replay_mode;   // mode of operation
 	bool log_reads;   // additionally log pre-images
+	bool cease_logging; // direct IO without logging (only in case of EMERGENCY)
 	bool debug_shortcut; // only for testing! never use in production!
 	loff_t replay_start_pos; // where to start replay
 	loff_t replay_end_pos;   // end of replay

sy_old/mars_light.c

@@ -87,6 +87,7 @@ struct mars_rotate {
 	struct mars_dent *next_next_relevant_log;
 	struct mars_dent *prev_log;
 	struct mars_dent *next_log;
+	struct mars_dent *syncstatus_dent;
 	struct if_brick *if_brick;
 	const char *copy_path;
 	struct copy_brick *copy_brick;
@@ -165,6 +166,8 @@ EXPORT_SYMBOL_GPL(mars_mem_percent);
 #define EXHAUSTED(x,max) (false)
 #endif
 
+#define JAMMED(x) ((x) <= 1024 * 1024)
+
 static
 int _set_trans_params(struct mars_brick *_brick, void *private)
 {
@@ -2348,6 +2351,40 @@ done:
 	return status;
 }
 
+static
+void override_all_syncstatus(struct mars_global *global, struct mars_rotate *rot, char *parent_path)
+{
+	char *prefix = NULL;
+	struct mars_dent **table = NULL;
+	int count;
+	int i;
+
+
+	prefix = path_make("%s/syncstatus-", parent_path);
+	if (!prefix)
+		goto done;
+		
+	count = mars_find_dent_all(global, prefix, &table);
+	MARS_DBG("prefix='%s' count=%d\n", prefix, count);
+
+	for (i = 0; i < count; i++) {
+		struct mars_dent *dent = table[i];
+		int status;
+
+		if (!dent || !dent->d_path || dent == rot->syncstatus_dent)
+			continue;
+		
+		status = mars_symlink("0", dent->d_path, NULL, 0);
+		MARS_DBG("clearing syncstatus link '%s' status=%d\n", dent->d_path, status);
+	}
+
+done:
+	if (table)
+		brick_mem_free(table);
+	if (prefix)
+		brick_string_free(prefix);
+}
+
 static
 int make_log_finalize(struct mars_global *global, struct mars_dent *dent)
 {
@@ -2367,6 +2404,17 @@ int make_log_finalize(struct mars_global *global, struct mars_dent *dent)
 		goto done;
 	}
 
+	/* Handle jamming (a very exceptional state)
+	 */
+	if (global->jammed) {
+		if (rot->todo_primary || rot->is_primary)
+			trans_brick->cease_logging = true;
+	} else if (!rot->todo_primary && !rot->is_primary) {
+		trans_brick->cease_logging = false;
+	}
+	if (trans_brick->cease_logging)
+		override_all_syncstatus(global, rot, parent->d_path);
+
 	// check whether some copy has finished
 	copy_brick = (struct copy_brick*)mars_find_brick(global, &copy_brick_type, rot->copy_path);
 	MARS_DBG("copy_path = '%s' copy_brick = %p\n", rot->copy_path, copy_brick);
@@ -2824,6 +2872,7 @@ static int make_sync(void *buf, struct mars_dent *dent)
 	rot = dent->d_parent->d_private;
 	if (rot) {
 		rot->allow_update = true;
+		rot->syncstatus_dent = dent;
 	}
 
 	/* Sync necessary?
@@ -3039,6 +3088,7 @@ enum {
 	CL_PEERS,
 	CL_ALIVE,
 	CL_EXHAUSTED,
+	CL_JAMMED,
 	CL_REST_SPACE,
 	// resource definitions
 	CL_RESOURCE,
@@ -3166,6 +3216,12 @@ static const struct light_class light_classes[] = {
 		.cl_type = 'l',
 		.cl_father = CL_ROOT,
 	},
+	[CL_JAMMED] = {
+		.cl_name = "jammed-",
+		.cl_len = 7,
+		.cl_type = 'l',
+		.cl_father = CL_ROOT,
+	},
 
 	/* Directory containing all items of a resource
 	 */
@@ -3703,6 +3759,7 @@ static int light_thread(void *data)
 		int status;
 		loff_t rest_space;
 		bool exhausted;
+		bool jammed;
 
 		MARS_DBG("-------- NEW ROUND ---------\n");
 
@@ -3718,11 +3775,19 @@ static int light_thread(void *data)
 		_make_alivelink("alive", _global.global_power.button ? 1 : 0);
 
 		mars_remaining_space("/mars", &_global.total_space, &_global.remaining_space);
+
 		exhausted = EXHAUSTED(_global.remaining_space, _global.total_space);
 		_global.exhausted = exhausted;
 		_make_alivelink("exhausted", exhausted ? 1 : 0);
 		if (exhausted)
-			MARS_WRN("EXHAUSTED filesystem space = %lld, STOPPING IO\n", _global.remaining_space);
+			MARS_WRN("EXHAUSTED filesystem space = %lld\n", _global.remaining_space);
+
+		jammed = JAMMED(_global.remaining_space);
+		_global.jammed = jammed;
+		_make_alivelink("jammed", jammed ? 1 : 0);
+		if (jammed)
+			MARS_WRN("JAMMED filesystem space = %lld, STOPPING TRANSACTION LOGGING\n", _global.remaining_space);
+
 		rest_space = _global.remaining_space - EXHAUSTED_LIMIT(_global.total_space);
 		_make_alivelink("rest-space", rest_space);
 

sy_old/strategy.h

@@ -59,6 +59,7 @@ struct mars_global {
 	int deleted_min;
 	bool main_trigger;
 	bool exhausted;
+	bool jammed;
 };
 
 typedef int (*mars_dent_checker_fn)(struct mars_dent *parent, const char *name, int namlen, unsigned int d_type, int *prefix, int *serial);