[thirdparty/mdadm.git] / Monitor.c

/*
 * mdadm - manage Linux "md" devices aka RAID arrays.
 *
 * Copyright (C) 2001-2002 Neil Brown <neilb@cse.unsw.edu.au>
 *
 *
 *    This program is free software; you can redistribute it and/or modify
 *    it under the terms of the GNU General Public License as published by
 *    the Free Software Foundation; either version 2 of the License, or
 *    (at your option) any later version.
 *
 *    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 for more details.
 *
 *    You should have received a copy of the GNU General Public License
 *    along with this program; if not, write to the Free Software
 *    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 *    Author: Neil Brown
 *    Email: <neilb@cse.unsw.edu.au>
 *    Paper: Neil Brown
 *           School of Computer Science and Engineering
 *           The University of New South Wales
 *           Sydney, 2052
 *           Australia
 */

#include	"mdadm.h"
#include	"md_p.h"
#include	"md_u.h"
#include	<sys/wait.h>
#include	<sys/signal.h>

static void alert(char *event, char *dev, char *disc, char *mailaddr, char *cmd);

static char *percentalerts[] = { 
	"RebuildStarted",
	"Rebuild20",
	"Rebuild40",
	"Rebuild60",
	"Rebuild80",
};

int Monitor(mddev_dev_t devlist,
	    char *mailaddr, char *alert_cmd,
	    int period, int daemonise, int scan, int oneshot,
	    char *config)
{
	/*
	 * Every few seconds, scan every md device looking for changes
	 * When a change is found, log it, possibly run the alert command,
	 * and possibly send Email
	 *
	 * For each array, we record:
	 *   Update time
	 *   active/working/failed/spare drives
	 *   State of each device.
	 *   %rebuilt if rebuilding
	 *
	 * If the update time changes, check out all the data again
	 * It is possible that we cannot get the state of each device
	 * due to bugs in the md kernel module.
	 * We also read /proc/mdstat to get rebuild percent,
	 * and to get state on all active devices incase of kernel bug.
	 *
	 * Events are:
	 *    Fail
	 *	An active device had Faulty set or Active/Sync removed
	 *    FailSpare
	 *      A spare device had Faulty set
	 *    SpareActive
	 *      An active device had a reverse transition
	 *    RebuildStarted
	 *      percent went from -1 to +ve
	 *    Rebuild20 Rebuild40 Rebuild60 Rebuild80
	 *      percent went from below to not-below that number
	 *    DeviceDisappeared
	 *      Couldn't access a device which was previously visible
	 *
	 * if we detect an array with active<raid and spare==0
	 * we look at other arrays that have same spare-group
	 * If we find one with active==raid and spare>0,
	 *  and if we can get_disk_info and find a name
	 *  Then we hot-remove and hot-add to the other array
	 *
	 * If devlist is NULL, then we can monitor everything because --scan
	 * was given.  We get an initial list from config file and add anything
	 * that appears in /proc/mdstat
	 */

	struct state {
		char *devname;
		int devnum;	/* to sync with mdstat info */
		long utime;
		int err;
		char *spare_group;
		int active, working, failed, spare, raid;
		int devstate[MD_SB_DISKS];
		int devid[MD_SB_DISKS];
		int percent;
		struct state *next;
	} *statelist = NULL;
	int finished = 0;
	struct mdstat_ent *mdstat = NULL;

	if (!mailaddr) {
		mailaddr = conf_get_mailaddr(config);
		if (mailaddr && ! scan)
			fprintf(stderr, Name ": Monitor using email address \"%s\" from config file\n",
			       mailaddr);
	}
	if (!alert_cmd) {
		alert_cmd = conf_get_program(config);
		if (alert_cmd && ! scan)
			fprintf(stderr, Name ": Monitor using program \"%s\" from config file\n",
			       alert_cmd);
	}
	if (scan && !mailaddr && !alert_cmd) {
		fprintf(stderr, Name ": No mail address or alert command - not monitoring.\n");
		return 1;
	}

	if (daemonise) {
		int pid = fork();
		if (pid > 0) {
			printf("%d\n", pid);
			return 0;
		}
		if (pid < 0) {
			perror("daemonise");
			return 1;
		}
		close(0);
		open("/dev/null", 3);
		dup2(0,1);
		dup2(0,2);
		setsid();
	}

	if (devlist == NULL) {
		mddev_ident_t mdlist = conf_get_ident(config, NULL);
		for (; mdlist; mdlist=mdlist->next) {
			struct state *st = malloc(sizeof *st);
			if (st == NULL)
				continue;
			st->devname = strdup(mdlist->devname);
			st->utime = 0;
			st->next = statelist;
			st->err = 0;
			st->devnum = -1;
			st->percent = -2;
			if (mdlist->spare_group)
				st->spare_group = strdup(mdlist->spare_group);
			else
				st->spare_group = NULL;
			statelist = st;
		}
	} else {
		mddev_dev_t dv;
		for (dv=devlist ; dv; dv=dv->next) {
			struct state *st = malloc(sizeof *st);
			if (st == NULL)
				continue;
			st->devname = strdup(dv->devname);
			st->utime = 0;
			st->next = statelist;
			st->err = 0;
			st->devnum = -1;
			st->percent = -2;
			st->spare_group = NULL;
			statelist = st;
		}
	}


	while (! finished) {
		int new_found = 0;
		struct state *st;

		if (mdstat)
			free_mdstat(mdstat);
		mdstat = mdstat_read();

		for (st=statelist; st; st=st->next) {
			mdu_array_info_t array;
			struct mdstat_ent *mse;
			char *dev = st->devname;
			int fd;
			int i;

			fd = open(dev, O_RDONLY);
			if (fd < 0) {
				if (!st->err)
					alert("DeviceDisappeared", dev, NULL,
					      mailaddr, alert_cmd);
/*					fprintf(stderr, Name ": cannot open %s: %s\n",
						dev, strerror(errno));
*/				st->err=1;
				continue;
			}
			if (ioctl(fd, GET_ARRAY_INFO, &array)<0) {
				if (!st->err)
					alert("DeviceDisappeared", dev, NULL,
					      mailaddr, alert_cmd);
/*					fprintf(stderr, Name ": cannot get array info for %s: %s\n",
						dev, strerror(errno));
*/				st->err=1;
				close(fd);
				continue;
			}
			if (array.level != 1 && array.level != 5 && array.level != -4) {
				if (!st->err)
					alert("DeviceDisappeared", dev, "Wrong-Level",
					      mailaddr, alert_cmd);
				st->err = 1;
				close(fd);
				continue;
			}
			if (st->devnum < 0) {
				struct stat stb;
				if (fstat(fd, &stb) == 0 &&
				    (S_IFMT&stb.st_mode)==S_IFBLK)
					st->devnum = MINOR(stb.st_rdev);
			}

			for (mse = mdstat ; mse ; mse=mse->next)
				if (mse->devnum == st->devnum) {
					mse->devnum = -1; /* flag it as "used" */
					break;
				}

			if (st->utime == array.utime &&
			    st->failed == array.failed_disks &&
			    st->working == array.working_disks &&
			    st->spare == array.spare_disks &&
			    (mse == NULL  || (
				    mse->percent == st->percent
				    ))) {
				close(fd);
				st->err = 0;
				continue;
			}
			if (st->utime == 0 && /* new array */
			    mse &&	/* is in /proc/mdstat */
			    mse->pattern && strchr(mse->pattern, '_') /* degraded */
				)
				alert("DegradedArray", dev, NULL, mailaddr, alert_cmd);

			if (mse &&
			    st->percent == -1 && 
			    mse->percent >= 0)
				alert("RebuildStarted", dev, NULL, mailaddr, alert_cmd);
			if (mse &&
			    st->percent >= 0 &&
			    mse->percent >= 0 &&
			    (mse->percent / 20) > (st->percent / 20))
				alert(percentalerts[mse->percent/20],
				      dev, NULL, mailaddr, alert_cmd);

			if (mse)
				st->percent = mse->percent;
					
			for (i=0; i<MD_SB_DISKS; i++) {
				mdu_disk_info_t disc;
				int newstate=0;
				int change;
				char *dv = NULL;
				disc.number = i;
				if (ioctl(fd, GET_DISK_INFO, &disc)>= 0) {
					newstate = disc.state;
					dv = map_dev(disc.major, disc.minor);
				} else if (mse &&  mse->pattern && i < strlen(mse->pattern))
					switch(mse->pattern[i]) {
					case 'U': newstate = 6 /* ACTIVE/SYNC */; break;
					case '_': newstate = 0; break;
					}
				change = newstate ^ st->devstate[i];
				if (st->utime && change && !st->err) {
					if (i < array.raid_disks &&
					    (((newstate&change)&(1<<MD_DISK_FAULTY)) ||
					     ((st->devstate[i]&change)&(1<<MD_DISK_ACTIVE)) ||
					     ((st->devstate[i]&change)&(1<<MD_DISK_SYNC)))
						)
						alert("Fail", dev, dv, mailaddr, alert_cmd);
					else if (i>=array.raid_disks &&
						 (disc.major || disc.minor) &&
						 st->devid[i] == MKDEV(disc.major, disc.minor) &&
						 ((newstate&change)&(1<<MD_DISK_FAULTY))
						)
						alert("FailSpare", dev, dv, mailaddr, alert_cmd);
					else if (i < array.raid_disks &&
						 (((st->devstate[i]&change)&(1<<MD_DISK_FAULTY)) ||
						  ((newstate&change)&(1<<MD_DISK_ACTIVE)) ||
						  ((newstate&change)&(1<<MD_DISK_SYNC)))
						)
						alert("SpareActive", dev, dv, mailaddr, alert_cmd);
				}
				st->devstate[i] = disc.state;
				st->devid[i] = MKDEV(disc.major, disc.minor);
			}
			close(fd);
			st->active = array.active_disks;
			st->working = array.working_disks;
			st->spare = array.spare_disks;
			st->failed = array.failed_disks;
			st->utime = array.utime;
			st->raid = array.raid_disks;
			st->err = 0;
		}
		/* now check if there are any new devices found in mdstat */
		if (scan) {
			struct mdstat_ent *mse;
			for (mse=mdstat; mse; mse=mse->next) 
				if (mse->devnum >= 0 &&
				    (strcmp(mse->level, "raid1")==0 ||
				     strcmp(mse->level, "raid5")==0 ||
				     strcmp(mse->level, "multipath")==0)
					) {
					struct state *st = malloc(sizeof *st);
					if (st == NULL)
						continue;
					st->devname = strdup(get_md_name(mse->devnum));
					st->utime = 0;
					st->next = statelist;
					st->err = 1;
					st->devnum = mse->devnum;
					st->percent = -2;
					st->spare_group = NULL;
					statelist = st;
					alert("NewArray", st->devname, NULL, mailaddr, alert_cmd);
					new_found = 1;
				}
		}
		/* If an array has active < raid && spare == 0 && spare_group != NULL
		 * Look for another array with spare > 0 and active == raid and same spare_group
		 *  if found, choose a device and hotremove/hotadd
		 */
		for (st = statelist; st; st=st->next)
			if (st->active < st->raid &&
			    st->spare == 0 &&
			    st->spare_group != NULL) {
				struct state *st2;
				for (st2=statelist ; st2 ; st2=st2->next)
					if (st2 != st &&
					    st2->spare > 0 &&
					    st2->active == st2->raid &&
					    st2->spare_group != NULL &&
					    strcmp(st->spare_group, st2->spare_group) == 0) {
						/* try to remove and add */
						int fd1 = open(st->devname, O_RDONLY);
						int fd2 = open(st2->devname, O_RDONLY);
						int dev = -1;
						int d;
						if (fd1 < 0 || fd2 < 0) {
							if (fd1>=0) close(fd1);
							if (fd2>=0) close(fd2);
							continue;
						}
						for (d=st2->raid; d<MD_SB_DISKS; d++) {
							if (st2->devid[d] > 0 &&
							    st2->devstate[d] == 0) {
								dev = st2->devid[d];
								break;
							}
						}
						if (dev > 0) {
							if (ioctl(fd2, HOT_REMOVE_DISK, 
								  (unsigned long)dev) == 0) {
								if (ioctl(fd1, HOT_ADD_DISK,
									  (unsigned long)dev) == 0) {
									alert("MoveSpare", st->devname, st2->devname, mailaddr, alert_cmd);
									close(fd1);
									close(fd2);
									break;
								}
								else ioctl(fd2, HOT_ADD_DISK, (unsigned long) dev);
							}
						}
						close(fd1);
						close(fd2);
					}
			}
		if (!new_found) {
			if (oneshot)
				break;
			else
				sleep(period);
		}
	}
	return 0;
}


static void alert(char *event, char *dev, char *disc, char *mailaddr, char *cmd)
{
	if (!cmd && !mailaddr) {
		time_t now = time(0);
	       
		printf("%1.15s: %s on %s %s\n", ctime(&now)+4, event, dev, disc?disc:"unknown device");
	}
	if (cmd) {
		int pid = fork();
		switch(pid) {
		default:
			waitpid(pid, NULL, 0);
			break;
		case -1:
			break;
		case 0:
			execl(cmd, cmd, event, dev, disc, NULL);
			exit(2);
		}
	}
	if (mailaddr && 
	    (strncmp(event, "Fail", 4)==0 || 
	     strncmp(event, "Degrade", 7)==0)) {
		FILE *mp = popen(Sendmail, "w");
		if (mp) {
			char hname[256];
			gethostname(hname, sizeof(hname));
			signal(SIGPIPE, SIG_IGN);
			fprintf(mp, "From: " Name " monitoring <root>\n");
			fprintf(mp, "To: %s\n", mailaddr);
			fprintf(mp, "Subject: %s event on %s:%s\n\n", event, dev, hname);

			fprintf(mp, "This is an automatically generated mail message from " Name "\n");
			fprintf(mp, "running on %s\n\n", hname);

			fprintf(mp, "A %s event had been detected on md device %s.\n\n", event, dev);

			if (disc)
				fprintf(mp, "It could be related to component device %s.\n\n", disc);

			fprintf(mp, "Faithfully yours, etc.\n");
			fclose(mp);
		}

	}
	/* FIXME log the event to syslog maybe */
}
Commit	Line	Data
52826846	1	/*
9a9dab36	2	* mdadm - manage Linux "md" devices aka RAID arrays.
52826846	3	*
cd29a5c8	4	* Copyright (C) 2001-2002 Neil Brown <neilb@cse.unsw.edu.au>
52826846 NB	5	*
	6	*
	7	* This program is free software; you can redistribute it and/or modify
	8	* it under the terms of the GNU General Public License as published by
	9	* the Free Software Foundation; either version 2 of the License, or
	10	* (at your option) any later version.
	11	*
	12	* This program is distributed in the hope that it will be useful,
	13	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	14	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	15	* GNU General Public License for more details.
	16	*
	17	* You should have received a copy of the GNU General Public License
	18	* along with this program; if not, write to the Free Software
	19	* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	20	*
	21	* Author: Neil Brown
	22	* Email: <neilb@cse.unsw.edu.au>
	23	* Paper: Neil Brown
	24	* School of Computer Science and Engineering
	25	* The University of New South Wales
	26	* Sydney, 2052
	27	* Australia
	28	*/
	29
9a9dab36	30	#include "mdadm.h"
52826846 NB	31	#include "md_p.h"
52826846 NB	32	#include "md_u.h"
e0d19036	33	#include <sys/wait.h>
52826846 NB	34	#include <sys/signal.h>
	35
	36	static void alert(char event, char dev, char disc, char mailaddr, char *cmd);
	37
e0d19036 NB	38	static char *percentalerts[] = {
	39	"RebuildStarted",
	40	"Rebuild20",
	41	"Rebuild40",
	42	"Rebuild60",
	43	"Rebuild80",
	44	};
	45
cd29a5c8	46	int Monitor(mddev_dev_t devlist,
52826846	47	char mailaddr, char alert_cmd,
aa88f531	48	int period, int daemonise, int scan, int oneshot,
52826846 NB	49	char *config)
	50	{
	51	/*
	52	* Every few seconds, scan every md device looking for changes
	53	* When a change is found, log it, possibly run the alert command,
	54	* and possibly send Email
	55	*
	56	* For each array, we record:
	57	* Update time
	58	* active/working/failed/spare drives
	59	* State of each device.
e0d19036	60	* %rebuilt if rebuilding
52826846 NB	61	*
	62	* If the update time changes, check out all the data again
	63	* It is possible that we cannot get the state of each device
	64	* due to bugs in the md kernel module.
e0d19036 NB	65	* We also read /proc/mdstat to get rebuild percent,
e0d19036 NB	66	* and to get state on all active devices incase of kernel bug.
52826846	67	*
e0d19036 NB	68	* Events are:
	69	* Fail
	70	* An active device had Faulty set or Active/Sync removed
	71	* FailSpare
	72	* A spare device had Faulty set
	73	* SpareActive
	74	* An active device had a reverse transition
	75	* RebuildStarted
	76	* percent went from -1 to +ve
	77	* Rebuild20 Rebuild40 Rebuild60 Rebuild80
	78	* percent went from below to not-below that number
	79	* DeviceDisappeared
	80	* Couldn't access a device which was previously visible
52826846 NB	81	*
	82	* if we detect an array with active<raid and spare==0
	83	* we look at other arrays that have same spare-group
	84	* If we find one with active==raid and spare>0,
	85	* and if we can get_disk_info and find a name
	86	* Then we hot-remove and hot-add to the other array
	87	*
e0d19036 NB	88	* If devlist is NULL, then we can monitor everything because --scan
	89	* was given. We get an initial list from config file and add anything
	90	* that appears in /proc/mdstat
52826846 NB	91	*/
	92
	93	struct state {
	94	char *devname;
e0d19036	95	int devnum; /* to sync with mdstat info */
52826846 NB	96	long utime;
52826846 NB	97	int err;
e0d19036 NB	98	char *spare_group;
e0d19036 NB	99	int active, working, failed, spare, raid;
52826846	100	int devstate[MD_SB_DISKS];
e0d19036 NB	101	int devid[MD_SB_DISKS];
e0d19036 NB	102	int percent;
52826846 NB	103	struct state *next;
	104	} *statelist = NULL;
	105	int finished = 0;
e0d19036 NB	106	struct mdstat_ent *mdstat = NULL;
e0d19036 NB	107
d013a55e	108	if (!mailaddr) {
e0d19036	109	mailaddr = conf_get_mailaddr(config);
d013a55e	110	if (mailaddr && ! scan)
56eedc1a	111	fprintf(stderr, Name ": Monitor using email address \"%s\" from config file\n",
d013a55e NB	112	mailaddr);
	113	}
	114	if (!alert_cmd) {
e0d19036	115	alert_cmd = conf_get_program(config);
d013a55e	116	if (alert_cmd && ! scan)
56eedc1a	117	fprintf(stderr, Name ": Monitor using program \"%s\" from config file\n",
d013a55e NB	118	alert_cmd);
d013a55e NB	119	}
56eedc1a NB	120	if (scan && !mailaddr && !alert_cmd) {
56eedc1a NB	121	fprintf(stderr, Name ": No mail address or alert command - not monitoring.\n");
d013a55e	122	return 1;
56eedc1a	123	}
d013a55e NB	124
	125	if (daemonise) {
	126	int pid = fork();
	127	if (pid > 0) {
	128	printf("%d\n", pid);
	129	return 0;
	130	}
	131	if (pid < 0) {
	132	perror("daemonise");
	133	return 1;
	134	}
	135	close(0);
	136	open("/dev/null", 3);
	137	dup2(0,1);
	138	dup2(0,2);
	139	setsid();
	140	}
e0d19036 NB	141
	142	if (devlist == NULL) {
	143	mddev_ident_t mdlist = conf_get_ident(config, NULL);
	144	for (; mdlist; mdlist=mdlist->next) {
	145	struct state st = malloc(sizeof st);
	146	if (st == NULL)
	147	continue;
	148	st->devname = strdup(mdlist->devname);
	149	st->utime = 0;
	150	st->next = statelist;
d013a55e	151	st->err = 0;
e0d19036 NB	152	st->devnum = -1;
	153	st->percent = -2;
	154	if (mdlist->spare_group)
	155	st->spare_group = strdup(mdlist->spare_group);
	156	else
	157	st->spare_group = NULL;
	158	statelist = st;
	159	}
	160	} else {
cd29a5c8	161	mddev_dev_t dv;
e0d19036 NB	162	for (dv=devlist ; dv; dv=dv->next) {
	163	struct state st = malloc(sizeof st);
	164	if (st == NULL)
	165	continue;
	166	st->devname = strdup(dv->devname);
	167	st->utime = 0;
	168	st->next = statelist;
d013a55e	169	st->err = 0;
e0d19036 NB	170	st->devnum = -1;
	171	st->percent = -2;
	172	st->spare_group = NULL;
	173	statelist = st;
	174	}
	175	}
	176
	177
	178	while (! finished) {
aa88f531	179	int new_found = 0;
e0d19036 NB	180	struct state *st;
	181
	182	if (mdstat)
	183	free_mdstat(mdstat);
	184	mdstat = mdstat_read();
	185
	186	for (st=statelist; st; st=st->next) {
52826846	187	mdu_array_info_t array;
e0d19036 NB	188	struct mdstat_ent *mse;
e0d19036 NB	189	char *dev = st->devname;
52826846	190	int fd;
52826846	191	int i;
e0d19036	192
52826846 NB	193	fd = open(dev, O_RDONLY);
	194	if (fd < 0) {
	195	if (!st->err)
e0d19036 NB	196	alert("DeviceDisappeared", dev, NULL,
	197	mailaddr, alert_cmd);
	198	/* fprintf(stderr, Name ": cannot open %s: %s\n",
52826846	199	dev, strerror(errno));
e0d19036	200	*/ st->err=1;
52826846 NB	201	continue;
	202	}
	203	if (ioctl(fd, GET_ARRAY_INFO, &array)<0) {
	204	if (!st->err)
e0d19036 NB	205	alert("DeviceDisappeared", dev, NULL,
	206	mailaddr, alert_cmd);
	207	/* fprintf(stderr, Name ": cannot get array info for %s: %s\n",
52826846	208	dev, strerror(errno));
e0d19036	209	*/ st->err=1;
52826846 NB	210	close(fd);
	211	continue;
	212	}
5787fa49 NB	213	if (array.level != 1 && array.level != 5 && array.level != -4) {
	214	if (!st->err)
	215	alert("DeviceDisappeared", dev, "Wrong-Level",
	216	mailaddr, alert_cmd);
	217	st->err = 1;
	218	close(fd);
	219	continue;
	220	}
e0d19036 NB	221	if (st->devnum < 0) {
	222	struct stat stb;
	223	if (fstat(fd, &stb) == 0 &&
	224	(S_IFMT&stb.st_mode)==S_IFBLK)
	225	st->devnum = MINOR(stb.st_rdev);
	226	}
	227
	228	for (mse = mdstat ; mse ; mse=mse->next)
	229	if (mse->devnum == st->devnum) {
	230	mse->devnum = -1; /* flag it as "used" */
	231	break;
	232	}
	233
52826846	234	if (st->utime == array.utime &&
e0d19036 NB	235	st->failed == array.failed_disks &&
	236	st->working == array.working_disks &&
	237	st->spare == array.spare_disks &&
	238	(mse == NULL \|\| (
	239	mse->percent == st->percent
	240	))) {
52826846	241	close(fd);
e0d19036	242	st->err = 0;
52826846 NB	243	continue;
52826846 NB	244	}
aa88f531 NB	245	if (st->utime == 0 && /* new array */
	246	mse && /* is in /proc/mdstat */
	247	mse->pattern && strchr(mse->pattern, '_') /* degraded */
	248	)
	249	alert("DegradedArray", dev, NULL, mailaddr, alert_cmd);
	250
e0d19036 NB	251	if (mse &&
	252	st->percent == -1 &&
	253	mse->percent >= 0)
	254	alert("RebuildStarted", dev, NULL, mailaddr, alert_cmd);
	255	if (mse &&
	256	st->percent >= 0 &&
	257	mse->percent >= 0 &&
	258	(mse->percent / 20) > (st->percent / 20))
	259	alert(percentalerts[mse->percent/20],
	260	dev, NULL, mailaddr, alert_cmd);
	261
	262	if (mse)
	263	st->percent = mse->percent;
	264
	265	for (i=0; i<MD_SB_DISKS; i++) {
52826846	266	mdu_disk_info_t disc;
e0d19036 NB	267	int newstate=0;
	268	int change;
	269	char *dv = NULL;
52826846 NB	270	disc.number = i;
52826846 NB	271	if (ioctl(fd, GET_DISK_INFO, &disc)>= 0) {
e0d19036 NB	272	newstate = disc.state;
e0d19036 NB	273	dv = map_dev(disc.major, disc.minor);
5787fa49	274	} else if (mse && mse->pattern && i < strlen(mse->pattern))
e0d19036 NB	275	switch(mse->pattern[i]) {
	276	case 'U': newstate = 6 /* ACTIVE/SYNC */; break;
	277	case '_': newstate = 0; break;
52826846	278	}
e0d19036 NB	279	change = newstate ^ st->devstate[i];
	280	if (st->utime && change && !st->err) {
	281	if (i < array.raid_disks &&
	282	(((newstate&change)&(1<<MD_DISK_FAULTY)) \|\|
	283	((st->devstate[i]&change)&(1<<MD_DISK_ACTIVE)) \|\|
	284	((st->devstate[i]&change)&(1<<MD_DISK_SYNC)))
	285	)
	286	alert("Fail", dev, dv, mailaddr, alert_cmd);
	287	else if (i>=array.raid_disks &&
	288	(disc.major \|\| disc.minor) &&
	289	st->devid[i] == MKDEV(disc.major, disc.minor) &&
	290	((newstate&change)&(1<<MD_DISK_FAULTY))
	291	)
	292	alert("FailSpare", dev, dv, mailaddr, alert_cmd);
	293	else if (i < array.raid_disks &&
	294	(((st->devstate[i]&change)&(1<<MD_DISK_FAULTY)) \|\|
	295	((newstate&change)&(1<<MD_DISK_ACTIVE)) \|\|
	296	((newstate&change)&(1<<MD_DISK_SYNC)))
	297	)
	298	alert("SpareActive", dev, dv, mailaddr, alert_cmd);
52826846	299	}
e0d19036 NB	300	st->devstate[i] = disc.state;
e0d19036 NB	301	st->devid[i] = MKDEV(disc.major, disc.minor);
52826846 NB	302	}
	303	close(fd);
	304	st->active = array.active_disks;
	305	st->working = array.working_disks;
	306	st->spare = array.spare_disks;
	307	st->failed = array.failed_disks;
	308	st->utime = array.utime;
e0d19036 NB	309	st->raid = array.raid_disks;
e0d19036 NB	310	st->err = 0;
52826846	311	}
e0d19036 NB	312	/* now check if there are any new devices found in mdstat */
	313	if (scan) {
	314	struct mdstat_ent *mse;
	315	for (mse=mdstat; mse; mse=mse->next)
5787fa49 NB	316	if (mse->devnum >= 0 &&
	317	(strcmp(mse->level, "raid1")==0 \|\|
	318	strcmp(mse->level, "raid5")==0 \|\|
	319	strcmp(mse->level, "multipath")==0)
	320	) {
e0d19036 NB	321	struct state st = malloc(sizeof st);
	322	if (st == NULL)
	323	continue;
	324	st->devname = strdup(get_md_name(mse->devnum));
	325	st->utime = 0;
	326	st->next = statelist;
	327	st->err = 1;
	328	st->devnum = mse->devnum;
	329	st->percent = -2;
	330	st->spare_group = NULL;
	331	statelist = st;
	332	alert("NewArray", st->devname, NULL, mailaddr, alert_cmd);
aa88f531	333	new_found = 1;
e0d19036 NB	334	}
	335	}
	336	/* If an array has active < raid && spare == 0 && spare_group != NULL
	337	* Look for another array with spare > 0 and active == raid and same spare_group
	338	* if found, choose a device and hotremove/hotadd
	339	*/
	340	for (st = statelist; st; st=st->next)
	341	if (st->active < st->raid &&
	342	st->spare == 0 &&
	343	st->spare_group != NULL) {
	344	struct state *st2;
	345	for (st2=statelist ; st2 ; st2=st2->next)
	346	if (st2 != st &&
	347	st2->spare > 0 &&
	348	st2->active == st2->raid &&
	349	st2->spare_group != NULL &&
	350	strcmp(st->spare_group, st2->spare_group) == 0) {
	351	/* try to remove and add */
	352	int fd1 = open(st->devname, O_RDONLY);
	353	int fd2 = open(st2->devname, O_RDONLY);
	354	int dev = -1;
	355	int d;
	356	if (fd1 < 0 \|\| fd2 < 0) {
	357	if (fd1>=0) close(fd1);
	358	if (fd2>=0) close(fd2);
	359	continue;
	360	}
	361	for (d=st2->raid; d<MD_SB_DISKS; d++) {
	362	if (st2->devid[d] > 0 &&
	363	st2->devstate[d] == 0) {
	364	dev = st2->devid[d];
	365	break;
	366	}
	367	}
	368	if (dev > 0) {
	369	if (ioctl(fd2, HOT_REMOVE_DISK,
	370	(unsigned long)dev) == 0) {
	371	if (ioctl(fd1, HOT_ADD_DISK,
	372	(unsigned long)dev) == 0) {
	373	alert("MoveSpare", st->devname, st2->devname, mailaddr, alert_cmd);
	374	close(fd1);
	375	close(fd2);
	376	break;
	377	}
	378	else ioctl(fd2, HOT_ADD_DISK, (unsigned long) dev);
	379	}
	380	}
	381	close(fd1);
	382	close(fd2);
	383	}
	384	}
aa88f531 NB	385	if (!new_found) {
	386	if (oneshot)
	387	break;
	388	else
	389	sleep(period);
	390	}
52826846 NB	391	}
	392	return 0;
	393	}
	394
	395
	396	static void alert(char event, char dev, char disc, char mailaddr, char *cmd)
	397	{
cd29a5c8 NB	398	if (!cmd && !mailaddr) {
	399	time_t now = time(0);
	400
e0d19036	401	printf("%1.15s: %s on %s %s\n", ctime(&now)+4, event, dev, disc?disc:"unknown device");
cd29a5c8	402	}
52826846 NB	403	if (cmd) {
	404	int pid = fork();
	405	switch(pid) {
	406	default:
	407	waitpid(pid, NULL, 0);
	408	break;
	409	case -1:
	410	break;
	411	case 0:
	412	execl(cmd, cmd, event, dev, disc, NULL);
	413	exit(2);
	414	}
	415	}
aa88f531 NB	416	if (mailaddr &&
	417	(strncmp(event, "Fail", 4)==0 \|\|
	418	strncmp(event, "Degrade", 7)==0)) {
52826846 NB	419	FILE *mp = popen(Sendmail, "w");
	420	if (mp) {
	421	char hname[256];
	422	gethostname(hname, sizeof(hname));
	423	signal(SIGPIPE, SIG_IGN);
	424	fprintf(mp, "From: " Name " monitoring <root>\n");
	425	fprintf(mp, "To: %s\n", mailaddr);
	426	fprintf(mp, "Subject: %s event on %s:%s\n\n", event, dev, hname);
	427
	428	fprintf(mp, "This is an automatically generated mail message from " Name "\n");
	429	fprintf(mp, "running on %s\n\n", hname);
	430
	431	fprintf(mp, "A %s event had been detected on md device %s.\n\n", event, dev);
	432
	433	if (disc)
56eedc1a	434	fprintf(mp, "It could be related to component device %s.\n\n", disc);
52826846 NB	435
	436	fprintf(mp, "Faithfully yours, etc.\n");
	437	fclose(mp);
	438	}
	439
	440	}
	441	/* FIXME log the event to syslog maybe */
	442	}