summaryrefslogtreecommitdiffstats
path: root/fs/ocfs2/cluster
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ocfs2/cluster')
-rw-r--r--fs/ocfs2/cluster/heartbeat.c12
-rw-r--r--fs/ocfs2/cluster/nodemanager.c51
-rw-r--r--fs/ocfs2/cluster/nodemanager.h7
-rw-r--r--fs/ocfs2/cluster/quorum.c16
-rw-r--r--fs/ocfs2/cluster/tcp.c10
-rw-r--r--fs/ocfs2/cluster/tcp_internal.h4
6 files changed, 86 insertions, 14 deletions
diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
index c452d116b892..5c9890006708 100644
--- a/fs/ocfs2/cluster/heartbeat.c
+++ b/fs/ocfs2/cluster/heartbeat.c
@@ -78,7 +78,7 @@ static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type);
unsigned int o2hb_dead_threshold = O2HB_DEFAULT_DEAD_THRESHOLD;
-/* Only sets a new threshold if there are no active regions.
+/* Only sets a new threshold if there are no active regions.
*
* No locking or otherwise interesting code is required for reading
* o2hb_dead_threshold as it can't change once regions are active and
@@ -170,13 +170,14 @@ static void o2hb_write_timeout(struct work_struct *work)
mlog(ML_ERROR, "Heartbeat write timeout to device %s after %u "
"milliseconds\n", reg->hr_dev_name,
- jiffies_to_msecs(jiffies - reg->hr_last_timeout_start));
+ jiffies_to_msecs(jiffies - reg->hr_last_timeout_start));
o2quo_disk_timeout();
}
static void o2hb_arm_write_timeout(struct o2hb_region *reg)
{
- mlog(0, "Queue write timeout for %u ms\n", O2HB_MAX_WRITE_TIMEOUT_MS);
+ mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n",
+ O2HB_MAX_WRITE_TIMEOUT_MS);
cancel_delayed_work(&reg->hr_write_timeout_work);
reg->hr_last_timeout_start = jiffies;
@@ -623,7 +624,7 @@ static int o2hb_check_slot(struct o2hb_region *reg,
"seq %llu last %llu changed %u equal %u\n",
slot->ds_node_num, (long long)slot->ds_last_generation,
le32_to_cpu(hb_block->hb_cksum),
- (unsigned long long)le64_to_cpu(hb_block->hb_seq),
+ (unsigned long long)le64_to_cpu(hb_block->hb_seq),
(unsigned long long)slot->ds_last_time, slot->ds_changed_samples,
slot->ds_equal_samples);
@@ -874,7 +875,8 @@ static int o2hb_thread(void *data)
do_gettimeofday(&after_hb);
elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb);
- mlog(0, "start = %lu.%lu, end = %lu.%lu, msec = %u\n",
+ mlog(ML_HEARTBEAT,
+ "start = %lu.%lu, end = %lu.%lu, msec = %u\n",
before_hb.tv_sec, (unsigned long) before_hb.tv_usec,
after_hb.tv_sec, (unsigned long) after_hb.tv_usec,
elapsed_msec);
diff --git a/fs/ocfs2/cluster/nodemanager.c b/fs/ocfs2/cluster/nodemanager.c
index 7ee6188bc79a..c81142e3ef84 100644
--- a/fs/ocfs2/cluster/nodemanager.c
+++ b/fs/ocfs2/cluster/nodemanager.c
@@ -35,6 +35,10 @@
* cluster references throughout where nodes are looked up */
struct o2nm_cluster *o2nm_single_cluster = NULL;
+char *o2nm_fence_method_desc[O2NM_FENCE_METHODS] = {
+ "reset", /* O2NM_FENCE_RESET */
+ "panic", /* O2NM_FENCE_PANIC */
+};
struct o2nm_node *o2nm_get_node_by_num(u8 node_num)
{
@@ -579,6 +583,43 @@ static ssize_t o2nm_cluster_attr_reconnect_delay_ms_write(
return o2nm_cluster_attr_write(page, count,
&cluster->cl_reconnect_delay_ms);
}
+
+static ssize_t o2nm_cluster_attr_fence_method_read(
+ struct o2nm_cluster *cluster, char *page)
+{
+ ssize_t ret = 0;
+
+ if (cluster)
+ ret = sprintf(page, "%s\n",
+ o2nm_fence_method_desc[cluster->cl_fence_method]);
+ return ret;
+}
+
+static ssize_t o2nm_cluster_attr_fence_method_write(
+ struct o2nm_cluster *cluster, const char *page, size_t count)
+{
+ unsigned int i;
+
+ if (page[count - 1] != '\n')
+ goto bail;
+
+ for (i = 0; i < O2NM_FENCE_METHODS; ++i) {
+ if (count != strlen(o2nm_fence_method_desc[i]) + 1)
+ continue;
+ if (strncasecmp(page, o2nm_fence_method_desc[i], count - 1))
+ continue;
+ if (cluster->cl_fence_method != i) {
+ printk(KERN_INFO "ocfs2: Changing fence method to %s\n",
+ o2nm_fence_method_desc[i]);
+ cluster->cl_fence_method = i;
+ }
+ return count;
+ }
+
+bail:
+ return -EINVAL;
+}
+
static struct o2nm_cluster_attribute o2nm_cluster_attr_idle_timeout_ms = {
.attr = { .ca_owner = THIS_MODULE,
.ca_name = "idle_timeout_ms",
@@ -603,10 +644,19 @@ static struct o2nm_cluster_attribute o2nm_cluster_attr_reconnect_delay_ms = {
.store = o2nm_cluster_attr_reconnect_delay_ms_write,
};
+static struct o2nm_cluster_attribute o2nm_cluster_attr_fence_method = {
+ .attr = { .ca_owner = THIS_MODULE,
+ .ca_name = "fence_method",
+ .ca_mode = S_IRUGO | S_IWUSR },
+ .show = o2nm_cluster_attr_fence_method_read,
+ .store = o2nm_cluster_attr_fence_method_write,
+};
+
static struct configfs_attribute *o2nm_cluster_attrs[] = {
&o2nm_cluster_attr_idle_timeout_ms.attr,
&o2nm_cluster_attr_keepalive_delay_ms.attr,
&o2nm_cluster_attr_reconnect_delay_ms.attr,
+ &o2nm_cluster_attr_fence_method.attr,
NULL,
};
static ssize_t o2nm_cluster_show(struct config_item *item,
@@ -778,6 +828,7 @@ static struct config_group *o2nm_cluster_group_make_group(struct config_group *g
cluster->cl_reconnect_delay_ms = O2NET_RECONNECT_DELAY_MS_DEFAULT;
cluster->cl_idle_timeout_ms = O2NET_IDLE_TIMEOUT_MS_DEFAULT;
cluster->cl_keepalive_delay_ms = O2NET_KEEPALIVE_DELAY_MS_DEFAULT;
+ cluster->cl_fence_method = O2NM_FENCE_RESET;
ret = &cluster->cl_group;
o2nm_single_cluster = cluster;
diff --git a/fs/ocfs2/cluster/nodemanager.h b/fs/ocfs2/cluster/nodemanager.h
index c992ea0da4ad..09ea2d388bbb 100644
--- a/fs/ocfs2/cluster/nodemanager.h
+++ b/fs/ocfs2/cluster/nodemanager.h
@@ -33,6 +33,12 @@
#include <linux/configfs.h>
#include <linux/rbtree.h>
+enum o2nm_fence_method {
+ O2NM_FENCE_RESET = 0,
+ O2NM_FENCE_PANIC,
+ O2NM_FENCE_METHODS, /* Number of fence methods */
+};
+
struct o2nm_node {
spinlock_t nd_lock;
struct config_item nd_item;
@@ -58,6 +64,7 @@ struct o2nm_cluster {
unsigned int cl_idle_timeout_ms;
unsigned int cl_keepalive_delay_ms;
unsigned int cl_reconnect_delay_ms;
+ enum o2nm_fence_method cl_fence_method;
/* this bitmap is part of a hack for disk bitmap.. will go eventually. - zab */
unsigned long cl_nodes_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
diff --git a/fs/ocfs2/cluster/quorum.c b/fs/ocfs2/cluster/quorum.c
index bbacf7da48a4..639024033fce 100644
--- a/fs/ocfs2/cluster/quorum.c
+++ b/fs/ocfs2/cluster/quorum.c
@@ -74,8 +74,20 @@ static void o2quo_fence_self(void)
* threads can still schedule, etc, etc */
o2hb_stop_all_regions();
- printk("ocfs2 is very sorry to be fencing this system by restarting\n");
- emergency_restart();
+ switch (o2nm_single_cluster->cl_fence_method) {
+ case O2NM_FENCE_PANIC:
+ panic("*** ocfs2 is very sorry to be fencing this system by "
+ "panicing ***\n");
+ break;
+ default:
+ WARN_ON(o2nm_single_cluster->cl_fence_method >=
+ O2NM_FENCE_METHODS);
+ case O2NM_FENCE_RESET:
+ printk(KERN_ERR "*** ocfs2 is very sorry to be fencing this "
+ "system by restarting ***\n");
+ emergency_restart();
+ break;
+ };
}
/* Indicate that a timeout occured on a hearbeat region write. The
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index 334f231a422c..d8d0c65ac03c 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -485,7 +485,7 @@ static void o2net_set_nn_state(struct o2net_node *nn,
}
if (was_valid && !valid) {
- printk(KERN_INFO "o2net: no longer connected to "
+ printk(KERN_NOTICE "o2net: no longer connected to "
SC_NODEF_FMT "\n", SC_NODEF_ARGS(old_sc));
o2net_complete_nodes_nsw(nn);
}
@@ -493,7 +493,7 @@ static void o2net_set_nn_state(struct o2net_node *nn,
if (!was_valid && valid) {
o2quo_conn_up(o2net_num_from_nn(nn));
cancel_delayed_work(&nn->nn_connect_expired);
- printk(KERN_INFO "o2net: %s " SC_NODEF_FMT "\n",
+ printk(KERN_NOTICE "o2net: %s " SC_NODEF_FMT "\n",
o2nm_this_node() > sc->sc_node->nd_num ?
"connected to" : "accepted connection from",
SC_NODEF_ARGS(sc));
@@ -930,7 +930,7 @@ static void o2net_sendpage(struct o2net_sock_container *sc,
cond_resched();
continue;
}
- mlog(ML_ERROR, "sendpage of size %zu to " SC_NODEF_FMT
+ mlog(ML_ERROR, "sendpage of size %zu to " SC_NODEF_FMT
" failed with %zd\n", size, SC_NODEF_ARGS(sc), ret);
o2net_ensure_shutdown(nn, sc, 0);
break;
@@ -1476,14 +1476,14 @@ static void o2net_idle_timer(unsigned long data)
do_gettimeofday(&now);
- printk(KERN_INFO "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u "
+ printk(KERN_NOTICE "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u "
"seconds, shutting it down.\n", SC_NODEF_ARGS(sc),
o2net_idle_timeout() / 1000,
o2net_idle_timeout() % 1000);
mlog(ML_NOTICE, "here are some times that might help debug the "
"situation: (tmr %ld.%ld now %ld.%ld dr %ld.%ld adv "
"%ld.%ld:%ld.%ld func (%08x:%u) %ld.%ld:%ld.%ld)\n",
- sc->sc_tv_timer.tv_sec, (long) sc->sc_tv_timer.tv_usec,
+ sc->sc_tv_timer.tv_sec, (long) sc->sc_tv_timer.tv_usec,
now.tv_sec, (long) now.tv_usec,
sc->sc_tv_data_ready.tv_sec, (long) sc->sc_tv_data_ready.tv_usec,
sc->sc_tv_advance_start.tv_sec,
diff --git a/fs/ocfs2/cluster/tcp_internal.h b/fs/ocfs2/cluster/tcp_internal.h
index 8d58cfe410b1..96fa7ebc530c 100644
--- a/fs/ocfs2/cluster/tcp_internal.h
+++ b/fs/ocfs2/cluster/tcp_internal.h
@@ -32,10 +32,10 @@
* on their number */
#define O2NET_QUORUM_DELAY_MS ((o2hb_dead_threshold + 2) * O2HB_REGION_TIMEOUT_MS)
-/*
+/*
* This version number represents quite a lot, unfortunately. It not
* only represents the raw network message protocol on the wire but also
- * locking semantics of the file system using the protocol. It should
+ * locking semantics of the file system using the protocol. It should
* be somewhere else, I'm sure, but right now it isn't.
*
* With version 11, we separate out the filesystem locking portion. The
OpenPOWER on IntegriCloud