[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

[Cluster-devel] cluster/cmirror-kernel/src dm-cmirror-client.c ...



CVSROOT:	/cvs/cluster
Module name:	cluster
Branch: 	RHEL4U4
Changes by:	jbrassow sourceware org	2006-07-27 23:10:58

Modified files:
	cmirror-kernel/src: dm-cmirror-client.c dm-cmirror-common.h 
	                    dm-cmirror-server.c 

Log message:
	- further tightening for 199826
	
	We now refuse to read/write the disk log if we are suspended.  We also
	add BUG_ON(<suspended>) to operations that do I/O to the log device.
	
	The reason for the BUG_ON() is that it is better to drop the machine
	than to have it hang the cluster while it attempts to read/write from
	a suspended device.  That being said, it should now be impossible to
	get to those functions which would perform I/O operations during
	suspension.
	
	I have still seen cases where the mirror will stall.  However, I think
	this is due to LVM (clvmd), because it happens when a mirror is created
	while the log device is suspended - which must not happen.  I've only
	seen this when doing simultaneous create/convert/remove from all nodes
	in the cluster.

Patches:
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/cmirror-kernel/src/dm-cmirror-client.c.diff?cvsroot=cluster&only_with_tag=RHEL4U4&r1=1.1.2.19.2.3&r2=1.1.2.19.2.4
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/cmirror-kernel/src/dm-cmirror-common.h.diff?cvsroot=cluster&only_with_tag=RHEL4U4&r1=1.1.2.7.2.2&r2=1.1.2.7.2.3
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/cmirror-kernel/src/dm-cmirror-server.c.diff?cvsroot=cluster&only_with_tag=RHEL4U4&r1=1.1.2.9.2.6&r2=1.1.2.9.2.7

--- cluster/cmirror-kernel/src/Attic/dm-cmirror-client.c	2006/07/22 22:12:32	1.1.2.19.2.3
+++ cluster/cmirror-kernel/src/Attic/dm-cmirror-client.c	2006/07/27 23:10:58	1.1.2.19.2.4
@@ -302,7 +302,7 @@
 		lc->server_id = lr.u.lr_coordinator;
 	} else {
 		/* ATTENTION -- what do we do with this ? */
-		DMWARN("Failed to receive election results from server");
+		DMWARN("Failed to receive election results from server: %d", len);
 		error = len;
 	}
 
@@ -363,21 +363,7 @@
 
 	iov.iov_len = sizeof(struct log_request);
 	iov.iov_base = lr;
-/*
-	DMERR("To  :: 0x%x, %s", 
-	       saddr_in.sin_addr.s_addr,
-	       (lr->lr_type == LRT_IS_CLEAN)? "LRT_IS_CLEAN":
-	       (lr->lr_type == LRT_IN_SYNC)? "LRT_IN_SYNC":
-	       (lr->lr_type == LRT_MARK_REGION)? "LRT_MARK_REGION":
-	       (lr->lr_type == LRT_GET_RESYNC_WORK)? "LRT_GET_RESYNC_WORK":
-	       (lr->lr_type == LRT_GET_SYNC_COUNT)? "LRT_GET_SYNC_COUNT":
-	       (lr->lr_type == LRT_CLEAR_REGION)? "LRT_CLEAR_REGION":
-	       (lr->lr_type == LRT_COMPLETE_RESYNC_WORK)? "LRT_COMPLETE_RESYNC_WORK":
-	       (lr->lr_type == LRT_MASTER_LEAVING)? "LRT_MASTER_LEAVING":
-	       (lr->lr_type == LRT_ELECTION)? "LRT_ELECTION":
-	       (lr->lr_type == LRT_SELECTION)? "LRT_SELECTION": "UNKNOWN"
-		);
-*/
+
 	if(lr->lr_type == LRT_MARK_REGION){
 		mark_req2ser++;
 	}
@@ -453,25 +439,28 @@
 			       request_retry_count,
 			       request_count,
 			       dm_div_up(request_retry_count*100, request_count));
+			DMDEBUG("Last request:");
+			DMDEBUG(" - my_id   :: %u", my_id);
+			DMDEBUG(" - server  :: %u", lc->server_id);
+			DMDEBUG(" - log uuid:: %s (%s)",
+			       lc->uuid + (strlen(lc->uuid) - 8),
+			       atomic_read(&lc->suspended) ? "suspended" : "active");
+			DMDEBUG(" - request :: %s",
+			       (type == LRT_IS_CLEAN)? "LRT_IS_CLEAN":
+			       (type == LRT_IN_SYNC)? "LRT_IN_SYNC":
+			       (type == LRT_MARK_REGION)? "LRT_MARK_REGION":
+			       (type == LRT_GET_RESYNC_WORK)? "LRT_GET_RESYNC_WORK":
+			       (type == LRT_GET_SYNC_COUNT)? "LRT_GET_SYNC_COUNT":
+			       (type == LRT_CLEAR_REGION)? "LRT_CLEAR_REGION":
+			       (type == LRT_COMPLETE_RESYNC_WORK)? "LRT_COMPLETE_RESYNC_WORK":
+			       (type == LRT_MASTER_LEAVING)? "LRT_MASTER_LEAVING":
+			       (type == LRT_ELECTION)? "LRT_ELECTION":
+			       (type == LRT_SELECTION)? "LRT_SELECTION": "UNKNOWN");
+			DMDEBUG(" - error   :: %d", error);
 		}
 	}
 
 	if(lr) kfree(lr);
-#if 0
-	DMINFO("My (%u) request (%s) to server (%u) failed :: %d",
-	       my_id,
-	       (type == LRT_IS_CLEAN)? "LRT_IS_CLEAN":
-	       (type == LRT_IN_SYNC)? "LRT_IN_SYNC":
-	       (type == LRT_MARK_REGION)? "LRT_MARK_REGION":
-	       (type == LRT_GET_RESYNC_WORK)? "LRT_GET_RESYNC_WORK":
-	       (type == LRT_GET_SYNC_COUNT)? "LRT_GET_SYNC_COUNT":
-	       (type == LRT_CLEAR_REGION)? "LRT_CLEAR_REGION":
-	       (type == LRT_COMPLETE_RESYNC_WORK)? "LRT_COMPLETE_RESYNC_WORK":
-	       (type == LRT_MASTER_LEAVING)? "LRT_MASTER_LEAVING":
-	       (type == LRT_ELECTION)? "LRT_ELECTION":
-	       (type == LRT_SELECTION)? "LRT_SELECTION": "UNKNOWN",
-	       lc->server_id, error);
-#endif
 	return error;
 }
 
@@ -739,7 +728,7 @@
 		DMINFO("Leaving while clear region requests remain.");
 
 	list_del_init(&lc->log_list);
-	if(lc->server_id == my_id)
+	if ((lc->server_id == my_id) && !atomic_read(&lc->suspended))
 		consult_server(lc, 0, LRT_MASTER_LEAVING, NULL);
 	sock_release(lc->client_sock);
 
--- cluster/cmirror-kernel/src/Attic/dm-cmirror-common.h	2006/07/22 22:12:32	1.1.2.7.2.2
+++ cluster/cmirror-kernel/src/Attic/dm-cmirror-common.h	2006/07/27 23:10:58	1.1.2.7.2.3
@@ -22,6 +22,7 @@
 #define DMWARN(f, x...) printk(KERN_WARNING DM_NAME ": " f "\n" , ## x)
 #define DMERR(f, x...) printk(KERN_ERR DM_NAME ": " f "\n" , ## x)
 #define DMINFO(f, x...) printk(KERN_INFO DM_NAME ": " f "\n" , ## x)
+#define DMDEBUG(f, x...) printk(KERN_DEBUG DM_NAME ": " f "\n" , ## x)
 #define DMEMIT(x...) sz += ((sz >= maxlen) ? \
 	  0 : scnprintf(result + sz, maxlen - sz, x))
 
--- cluster/cmirror-kernel/src/Attic/dm-cmirror-server.c	2006/07/22 22:49:49	1.1.2.9.2.6
+++ cluster/cmirror-kernel/src/Attic/dm-cmirror-server.c	2006/07/27 23:10:58	1.1.2.9.2.7
@@ -107,6 +107,7 @@
 	if (!log->log_dev)
 		return 0;
 
+	BUG_ON(atomic_read(&log->suspended));
 	r = dm_io_sync_vm(1, &log->header_location, READ,
 			  log->disk_header, &ebits);
 	if (unlikely(r))
@@ -137,6 +138,7 @@
 	if (!log->log_dev)
 		return 0;
 
+	BUG_ON(atomic_read(&log->suspended));
 	header_to_disk(&log->header, log->disk_header);
 	return dm_io_sync_vm(1, &log->header_location, WRITE,
 			     log->disk_header, &ebits);
@@ -180,6 +182,7 @@
 	if (!log->log_dev)
 		return 0;
 
+	BUG_ON(atomic_read(&log->suspended));
 	r = dm_io_sync_vm(1, &log->bits_location, READ,
 			  log->clean_bits, &ebits);
 
@@ -196,6 +199,7 @@
 	if (!log->log_dev)
 		return 0;
 
+	BUG_ON(atomic_read(&log->suspended));
 	return dm_io_sync_vm(1, &log->bits_location, WRITE,
 			     log->clean_bits, &ebits);
 }
@@ -295,7 +299,8 @@
 	struct region_user *tmp_ru, *ru;
 	unsigned char live_nodes[16]; /* Attention -- max of 128 nodes... */
 
-	DMINFO("Disk Resume::");
+	DMDEBUG("Disk Resume::  %s (%s)", lc->uuid + (strlen(lc->uuid) - 8),
+		atomic_read(&lc->suspended) ? "suspended" : "active");
 
 	debug_disk_write = 1;
 	memset(live_nodes, 0, sizeof(live_nodes));
@@ -355,20 +360,20 @@
 		}
 	}
 
-	DMINFO("  Live nodes        :: %d", global_count);
-	DMINFO("  In-Use Regions    :: %d", good_count+bad_count);
-	DMINFO("  Good IUR's        :: %d", good_count);
-	DMINFO("  Bad IUR's         :: %d", bad_count);
+	DMDEBUG("  Live nodes        :: %d", global_count);
+	DMDEBUG("  In-Use Regions    :: %d", good_count+bad_count);
+	DMDEBUG("  Good IUR's        :: %d", good_count);
+	DMDEBUG("  Bad IUR's         :: %d", bad_count);
 
 	lc->sync_count = count_bits32(lc->sync_bits, lc->bitset_uint32_count);
 	lc->sync_search = 0;
 
-	DMINFO("  Sync count        :: %Lu", lc->sync_count);
-	DMINFO("  Disk Region count :: %Lu", lc->header.nr_regions);
-	DMINFO("  Region count      :: %Lu", lc->region_count);
+	DMDEBUG("  Sync count        :: %Lu", lc->sync_count);
+	DMDEBUG("  Disk Region count :: %Lu", lc->header.nr_regions);
+	DMDEBUG("  Region count      :: %Lu", lc->region_count);
 
 	if(lc->header.nr_regions != lc->region_count){
-		DMINFO("  NOTE:  Mapping has changed.");
+		DMDEBUG("  NOTE:  Mapping has changed.");
 	}
 /* Take this out for now.
 	if(list_empty(&lc->region_users) && (lc->sync_count != lc->header.nr_regions)){
@@ -398,13 +403,13 @@
 	}			
 
 */
-	DMINFO("Marked regions::");
+	DMDEBUG("Marked regions::");
 	i = print_zero_bits((unsigned char *)lc->clean_bits, 0, lc->region_count);
-	DMINFO("  Total = %d", i);
+	DMDEBUG("  Total = %d", i);
 
-	DMINFO("Out-of-sync regions::");
+	DMDEBUG("Out-of-sync regions::");
 	i = print_zero_bits((unsigned char *)lc->sync_bits, 0, lc->region_count);
-	DMINFO("  Total = %d", i);
+	DMDEBUG("  Total = %d", i);
 
 	/* set the correct number of regions in the header */
 	lc->header.nr_regions = lc->region_count;
@@ -529,7 +534,7 @@
 
 	ru = find_ru(lc, who, lr->u.lr_region);
 	if(!ru){
-		DMINFO("Request to remove unrecorded region user (%u/%Lu)",
+		DMDEBUG("Request to remove unrecorded region user (%u/%Lu)",
 		       who, lr->u.lr_region);
 		return -EINVAL;
 	} else {
@@ -596,7 +601,7 @@
 	   (info < 1000 && !(info%100)) ||
 	   (info < 200 && !(info%25)) ||
 	   (info < 6)){
-		DMINFO(SECTOR_FORMAT " out-of-sync regions remaining for %s.",
+		DMDEBUG(SECTOR_FORMAT " out-of-sync regions remaining for %s.",
 		       lc->region_count - lc->sync_count,
 		       lc->uuid + (strlen(lc->uuid) - 8));
 	}
@@ -843,9 +848,12 @@
 				return -1;
 			}
 			if(lc && (old != lc->server_id) && (my_id == lc->server_id)){
-				DMINFO("I'm the cluster mirror log server for %s",
+				DMDEBUG("I'm the cluster mirror log server for %s",
 				       lc->uuid + (strlen(lc->uuid) - 8));
-				disk_resume(lc);
+				if (!atomic_read(&lc->suspended))
+					disk_resume(lc);
+				else
+					DMDEBUG("Not reading disk log because I'm suspended.");
 			}
 			goto reply;
 		}
@@ -860,6 +868,30 @@
 			goto reply;
 		}
 
+		if (atomic_read(&lc->suspended)) {
+			nodeid = ipaddr_to_nodeid((struct sockaddr *)msg.msg_name);
+			/*
+			DMDEBUG("Getting request while server (%u) is suspended:", my_id);
+			DMDEBUG(" - Requester :: %u", nodeid);
+			DMDEBUG(" - log uuid  :: %s", lc->uuid + (strlen(lc->uuid) - 8));
+			DMDEBUG(" - req type  :: %s",
+				(lr.lr_type == LRT_IS_CLEAN)? "LRT_IS_C	LEAN":
+				(lr.lr_type == LRT_IN_SYNC)? "LRT_IN_SYNC":
+				(lr.lr_type == LRT_MARK_REGION)? "LRT_MARK_REGION":
+				(lr.lr_type == LRT_GET_RESYNC_WORK)? "LRT_GET_RESYNC_WORK":
+				(lr.lr_type == LRT_GET_SYNC_COUNT)? "LRT_GET_SYNC_COUNT":
+				(lr.lr_type == LRT_CLEAR_REGION)? "LRT_CLEAR_REGION":
+				(lr.lr_type == LRT_COMPLETE_RESYNC_WORK)? "LRT_COMPLETE_RESYNC_WORK":
+				(lr.lr_type == LRT_MASTER_LEAVING)? "LRT_MASTER_LEAVING":
+				(lr.lr_type == LRT_ELECTION)? "LRT_ELECTION":
+				(lr.lr_type == LRT_SELECTION)? "LRT_SELECTION": "UNKNOWN");
+			*/
+			if (my_id != nodeid) {
+				lr.u.lr_int_rtn = -ENXIO;
+				goto reply;
+			}
+		}			
+
 		switch(lr.lr_type){
 		case LRT_IS_CLEAN:
 			error = server_is_clean(lc, &lr);
@@ -1004,7 +1036,10 @@
 			
 			list_for_each_entry(lc, &log_list_head, log_list){
 				if(lc->server_id == my_id){
-					disk_resume(lc);
+					if (!atomic_read(&lc->suspended))
+						disk_resume(lc);
+					else
+						DMDEBUG("Not reading disk log because I'm suspended.");
 				}
 			}
 			break;


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]