+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2011, Whamcloud, Inc.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
#include <stdio.h>
#include <errno.h>
#include <string.h>
timenow ()
{
struct timeval tv;
-
+
gettimeofday (&tv, NULL);
return (tv.tv_sec + tv.tv_usec / 1000000.0);
}
+typedef struct {
+ unsigned long msgs_alloc;
+ unsigned long msgs_max;
+ unsigned long errors;
+ unsigned long send_count;
+ unsigned long recv_count;
+ unsigned long route_count;
+ unsigned long drop_count;
+ unsigned long long send_length;
+ unsigned long long recv_length;
+ unsigned long long route_length;
+ unsigned long long drop_length;
+} counters_t;
+
+unsigned long long subull(unsigned long long a, unsigned long long b)
+{
+ if (a < b)
+ return -1ULL - b + a + 1;
+
+ return a - b;
+}
+
+unsigned long long subul(unsigned long a, unsigned long b)
+{
+ if (a < b)
+ return -1UL - b + a + 1;
+
+ return a - b;
+}
+
+double rul(unsigned long a, double secs)
+{
+ return (double)a/secs;
+}
+
+double rull(unsigned long long a, double secs)
+{
+ return (double)a/secs;
+}
+
void
do_stat (int fd)
{
static char buffer[1024];
static double last = 0.0;
- static unsigned long long old_bytes;
- static unsigned long old_packets;
- static unsigned long old_errors;
+ static counters_t old_counter;
double now;
double t;
- unsigned long long new_bytes, bytes;
- unsigned long new_packets, packets;
- unsigned long new_errors, errors;
- unsigned long depth;
+ counters_t new_counter;
+ counters_t counter;
int n;
-
+
lseek (fd, 0, SEEK_SET);
now = timenow();
n = read (fd, buffer, sizeof (buffer));
{
fprintf (stderr, "Can't read statfile\n");
exit (1);
- }
+ }
buffer[n] = 0;
-
- n = sscanf (buffer, "%Lu %lu %lu %lu",
- &new_bytes, &new_packets, &new_errors, &depth);
-
- if (n < 3)
+
+ n = sscanf(buffer, "%lu %lu %lu %lu %lu %lu %lu %llu %llu %llu %llu",
+ &new_counter.msgs_alloc, &new_counter.msgs_max,
+ &new_counter.errors,
+ &new_counter.send_count, &new_counter.recv_count,
+ &new_counter.route_count, &new_counter.drop_count,
+ &new_counter.send_length, &new_counter.recv_length,
+ &new_counter.route_length, &new_counter.drop_length);
+ if (n < 11)
{
fprintf (stderr, "Can't parse statfile\n");
exit (1);
}
-
- if (last == 0.0)
- printf ("%llu bytes, %lu packets (sz %lld), %lu errors",
- new_bytes, new_packets,
- (long long)((new_packets == 0) ? 0LL : new_bytes/new_packets),
- new_errors);
- else
- {
- t = now - last;
-
- if (new_bytes < old_bytes)
- bytes = -1ULL - old_bytes + new_bytes + 1;
- else
- bytes = new_bytes - old_bytes;
- if (new_packets < old_packets)
- packets = -1UL - old_packets + new_packets + 1;
- else
- packets = new_packets - old_packets;
- if (new_errors < old_errors)
- errors = -1UL - old_errors + new_errors + 1;
- else
- errors = new_errors - old_errors;
-
- printf ("%9llu bytes (%7.2fMb/s), %7lu packets (sz %5lld, %5ld/s), %lu errors (%ld/s)",
- bytes, ((double)bytes)/((1<<20) * t),
- packets, (long long)((packets == 0) ? 0LL : bytes/packets), (long)(packets/t),
- errors, (long)(errors/t));
- }
- old_bytes = new_bytes;
- old_packets = new_packets;
- old_errors = new_errors;
- if (n == 4)
- printf (", depth (%ld)\n", depth);
- else
- printf ("\n");
+ if (last == 0.0) {
+ printf("M %lu(%lu) E %lu S %llu/%lu R %llu/%lu F %llu/%lu "
+ "D %llu/%lu\n",
+ new_counter.msgs_alloc, new_counter.msgs_max,
+ new_counter.errors,
+ new_counter.send_length, new_counter.send_count,
+ new_counter.recv_length, new_counter.recv_count,
+ new_counter.route_length, new_counter.route_count,
+ new_counter.drop_length, new_counter.drop_count);
+ } else {
+ t = now - last;
+
+ counter.msgs_alloc = new_counter.msgs_alloc;
+ counter.msgs_max = new_counter.msgs_max;
+ counter.errors = subul(new_counter.errors, old_counter.errors);
+ counter.send_count = subul(new_counter.send_count, old_counter.send_count);
+ counter.recv_count = subul(new_counter.recv_count, old_counter.recv_count);
+ counter.route_count = subul(new_counter.route_count, old_counter.route_count);
+ counter.drop_count = subul(new_counter.drop_count, old_counter.drop_count);
+ counter.send_length = subull(new_counter.send_length, old_counter.send_length);
+ counter.recv_length = subull(new_counter.recv_length, old_counter.recv_length);
+ counter.route_length = subull(new_counter.route_length, old_counter.route_length);
+ counter.drop_length = subull(new_counter.drop_length, old_counter.drop_length);
+
+ printf ("M %3lu(%3lu) E %0.0f S %7.2f/%6.0f R %7.2f/%6.0f F %7.2f/%6.0f D %4.2f/%0.0f\n",
+ counter.msgs_alloc, counter.msgs_max,
+ rul(counter.errors,t),
+ rull(counter.send_length,t*1024.0*1024.0), rul(counter.send_count, t),
+ rull(counter.recv_length,t*1024.0*1024.0), rul(counter.recv_count, t),
+ rull(counter.route_length,t*1024.0*1024.0), rul(counter.route_count, t),
+ rull(counter.drop_length,t*1024.0*1024.0), rul(counter.drop_count, t));
+ }
+
+ old_counter = new_counter;
fflush (stdout);
-
+
lseek (fd, 0, SEEK_SET);
last = timenow();
}
{
int interval = 0;
int fd;
-
+
if (argc > 1)
interval = atoi (argv[1]);
- fd = open ("/proc/sys/portals/router", O_RDONLY);
+ fd = open ("/proc/sys/lnet/stats", O_RDONLY);
if (fd < 0)
{
fprintf (stderr, "Can't open stat: %s\n", strerror (errno));
return (1);
}
-
+
do_stat (fd);
if (interval == 0)
return (0);
-
+
for (;;)
{
sleep (interval);