[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 scan,
	    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 && scan)
		mailaddr = conf_get_mailaddr(config);
	if (!alert_cmd && scan)
		alert_cmd = conf_get_program(config);
	if (scan && !mailaddr && !alert_cmd)
		return 0;

	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 = 1;
			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 = 1;
			st->devnum = -1;
			st->percent = -2;
			st->spare_group = NULL;
			statelist = st;
		}
	}


	while (! finished) {
		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 (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);
				}
		}
		/* 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);
					}
			}
					    
		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) {
		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 componenet 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,
e0d19036	48	int period, int scan,
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;
	107
	108	if (!mailaddr && scan)
	109	mailaddr = conf_get_mailaddr(config);
	110	if (!alert_cmd && scan)
	111	alert_cmd = conf_get_program(config);
	112	if (scan && !mailaddr && !alert_cmd)
	113	return 0;
	114
	115	if (devlist == NULL) {
	116	mddev_ident_t mdlist = conf_get_ident(config, NULL);
	117	for (; mdlist; mdlist=mdlist->next) {
	118	struct state st = malloc(sizeof st);
	119	if (st == NULL)
	120	continue;
	121	st->devname = strdup(mdlist->devname);
	122	st->utime = 0;
	123	st->next = statelist;
	124	st->err = 1;
	125	st->devnum = -1;
	126	st->percent = -2;
	127	if (mdlist->spare_group)
	128	st->spare_group = strdup(mdlist->spare_group);
	129	else
	130	st->spare_group = NULL;
	131	statelist = st;
	132	}
	133	} else {
cd29a5c8	134	mddev_dev_t dv;
e0d19036 NB	135	for (dv=devlist ; dv; dv=dv->next) {
	136	struct state st = malloc(sizeof st);
	137	if (st == NULL)
	138	continue;
	139	st->devname = strdup(dv->devname);
	140	st->utime = 0;
	141	st->next = statelist;
	142	st->err = 1;
	143	st->devnum = -1;
	144	st->percent = -2;
	145	st->spare_group = NULL;
	146	statelist = st;
	147	}
	148	}
	149
	150
	151	while (! finished) {
	152	struct state *st;
	153
	154	if (mdstat)
	155	free_mdstat(mdstat);
	156	mdstat = mdstat_read();
	157
	158	for (st=statelist; st; st=st->next) {
52826846	159	mdu_array_info_t array;
e0d19036 NB	160	struct mdstat_ent *mse;
e0d19036 NB	161	char *dev = st->devname;
52826846	162	int fd;
52826846	163	int i;
e0d19036	164
52826846 NB	165	fd = open(dev, O_RDONLY);
	166	if (fd < 0) {
	167	if (!st->err)
e0d19036 NB	168	alert("DeviceDisappeared", dev, NULL,
	169	mailaddr, alert_cmd);
	170	/* fprintf(stderr, Name ": cannot open %s: %s\n",
52826846	171	dev, strerror(errno));
e0d19036	172	*/ st->err=1;
52826846 NB	173	continue;
	174	}
	175	if (ioctl(fd, GET_ARRAY_INFO, &array)<0) {
	176	if (!st->err)
e0d19036 NB	177	alert("DeviceDisappeared", dev, NULL,
	178	mailaddr, alert_cmd);
	179	/* fprintf(stderr, Name ": cannot get array info for %s: %s\n",
52826846	180	dev, strerror(errno));
e0d19036	181	*/ st->err=1;
52826846 NB	182	close(fd);
	183	continue;
	184	}
5787fa49 NB	185	if (array.level != 1 && array.level != 5 && array.level != -4) {
	186	if (!st->err)
	187	alert("DeviceDisappeared", dev, "Wrong-Level",
	188	mailaddr, alert_cmd);
	189	st->err = 1;
	190	close(fd);
	191	continue;
	192	}
e0d19036 NB	193	if (st->devnum < 0) {
	194	struct stat stb;
	195	if (fstat(fd, &stb) == 0 &&
	196	(S_IFMT&stb.st_mode)==S_IFBLK)
	197	st->devnum = MINOR(stb.st_rdev);
	198	}
	199
	200	for (mse = mdstat ; mse ; mse=mse->next)
	201	if (mse->devnum == st->devnum) {
	202	mse->devnum = -1; /* flag it as "used" */
	203	break;
	204	}
	205
52826846	206	if (st->utime == array.utime &&
e0d19036 NB	207	st->failed == array.failed_disks &&
	208	st->working == array.working_disks &&
	209	st->spare == array.spare_disks &&
	210	(mse == NULL \|\| (
	211	mse->percent == st->percent
	212	))) {
52826846	213	close(fd);
e0d19036	214	st->err = 0;
52826846 NB	215	continue;
52826846 NB	216	}
e0d19036 NB	217	if (mse &&
	218	st->percent == -1 &&
	219	mse->percent >= 0)
	220	alert("RebuildStarted", dev, NULL, mailaddr, alert_cmd);
	221	if (mse &&
	222	st->percent >= 0 &&
	223	mse->percent >= 0 &&
	224	(mse->percent / 20) > (st->percent / 20))
	225	alert(percentalerts[mse->percent/20],
	226	dev, NULL, mailaddr, alert_cmd);
	227
	228	if (mse)
	229	st->percent = mse->percent;
	230
	231	for (i=0; i<MD_SB_DISKS; i++) {
52826846	232	mdu_disk_info_t disc;
e0d19036 NB	233	int newstate=0;
	234	int change;
	235	char *dv = NULL;
52826846 NB	236	disc.number = i;
52826846 NB	237	if (ioctl(fd, GET_DISK_INFO, &disc)>= 0) {
e0d19036 NB	238	newstate = disc.state;
e0d19036 NB	239	dv = map_dev(disc.major, disc.minor);
5787fa49	240	} else if (mse && mse->pattern && i < strlen(mse->pattern))
e0d19036 NB	241	switch(mse->pattern[i]) {
	242	case 'U': newstate = 6 /* ACTIVE/SYNC */; break;
	243	case '_': newstate = 0; break;
52826846	244	}
e0d19036 NB	245	change = newstate ^ st->devstate[i];
	246	if (st->utime && change && !st->err) {
	247	if (i < array.raid_disks &&
	248	(((newstate&change)&(1<<MD_DISK_FAULTY)) \|\|
	249	((st->devstate[i]&change)&(1<<MD_DISK_ACTIVE)) \|\|
	250	((st->devstate[i]&change)&(1<<MD_DISK_SYNC)))
	251	)
	252	alert("Fail", dev, dv, mailaddr, alert_cmd);
	253	else if (i>=array.raid_disks &&
	254	(disc.major \|\| disc.minor) &&
	255	st->devid[i] == MKDEV(disc.major, disc.minor) &&
	256	((newstate&change)&(1<<MD_DISK_FAULTY))
	257	)
	258	alert("FailSpare", dev, dv, mailaddr, alert_cmd);
	259	else if (i < array.raid_disks &&
	260	(((st->devstate[i]&change)&(1<<MD_DISK_FAULTY)) \|\|
	261	((newstate&change)&(1<<MD_DISK_ACTIVE)) \|\|
	262	((newstate&change)&(1<<MD_DISK_SYNC)))
	263	)
	264	alert("SpareActive", dev, dv, mailaddr, alert_cmd);
52826846	265	}
e0d19036 NB	266	st->devstate[i] = disc.state;
e0d19036 NB	267	st->devid[i] = MKDEV(disc.major, disc.minor);
52826846 NB	268	}
	269	close(fd);
	270	st->active = array.active_disks;
	271	st->working = array.working_disks;
	272	st->spare = array.spare_disks;
	273	st->failed = array.failed_disks;
	274	st->utime = array.utime;
e0d19036 NB	275	st->raid = array.raid_disks;
e0d19036 NB	276	st->err = 0;
52826846	277	}
e0d19036 NB	278	/* now check if there are any new devices found in mdstat */
	279	if (scan) {
	280	struct mdstat_ent *mse;
	281	for (mse=mdstat; mse; mse=mse->next)
5787fa49 NB	282	if (mse->devnum >= 0 &&
	283	(strcmp(mse->level, "raid1")==0 \|\|
	284	strcmp(mse->level, "raid5")==0 \|\|
	285	strcmp(mse->level, "multipath")==0)
	286	) {
e0d19036 NB	287	struct state st = malloc(sizeof st);
	288	if (st == NULL)
	289	continue;
	290	st->devname = strdup(get_md_name(mse->devnum));
	291	st->utime = 0;
	292	st->next = statelist;
	293	st->err = 1;
	294	st->devnum = mse->devnum;
	295	st->percent = -2;
	296	st->spare_group = NULL;
	297	statelist = st;
	298	alert("NewArray", st->devname, NULL, mailaddr, alert_cmd);
	299	}
	300	}
	301	/* If an array has active < raid && spare == 0 && spare_group != NULL
	302	* Look for another array with spare > 0 and active == raid and same spare_group
	303	* if found, choose a device and hotremove/hotadd
	304	*/
	305	for (st = statelist; st; st=st->next)
	306	if (st->active < st->raid &&
	307	st->spare == 0 &&
	308	st->spare_group != NULL) {
	309	struct state *st2;
	310	for (st2=statelist ; st2 ; st2=st2->next)
	311	if (st2 != st &&
	312	st2->spare > 0 &&
	313	st2->active == st2->raid &&
	314	st2->spare_group != NULL &&
	315	strcmp(st->spare_group, st2->spare_group) == 0) {
	316	/* try to remove and add */
	317	int fd1 = open(st->devname, O_RDONLY);
	318	int fd2 = open(st2->devname, O_RDONLY);
	319	int dev = -1;
	320	int d;
	321	if (fd1 < 0 \|\| fd2 < 0) {
	322	if (fd1>=0) close(fd1);
	323	if (fd2>=0) close(fd2);
	324	continue;
	325	}
	326	for (d=st2->raid; d<MD_SB_DISKS; d++) {
	327	if (st2->devid[d] > 0 &&
	328	st2->devstate[d] == 0) {
	329	dev = st2->devid[d];
	330	break;
	331	}
	332	}
	333	if (dev > 0) {
	334	if (ioctl(fd2, HOT_REMOVE_DISK,
	335	(unsigned long)dev) == 0) {
	336	if (ioctl(fd1, HOT_ADD_DISK,
	337	(unsigned long)dev) == 0) {
	338	alert("MoveSpare", st->devname, st2->devname, mailaddr, alert_cmd);
	339	close(fd1);
	340	close(fd2);
	341	break;
	342	}
	343	else ioctl(fd2, HOT_ADD_DISK, (unsigned long) dev);
	344	}
	345	}
	346	close(fd1);
	347	close(fd2);
	348	}
	349	}
	350
52826846 NB	351	sleep(period);
	352	}
	353	return 0;
	354	}
	355
	356
	357	static void alert(char event, char dev, char disc, char mailaddr, char *cmd)
	358	{
cd29a5c8 NB	359	if (!cmd && !mailaddr) {
	360	time_t now = time(0);
	361
e0d19036	362	printf("%1.15s: %s on %s %s\n", ctime(&now)+4, event, dev, disc?disc:"unknown device");
cd29a5c8	363	}
52826846 NB	364	if (cmd) {
	365	int pid = fork();
	366	switch(pid) {
	367	default:
	368	waitpid(pid, NULL, 0);
	369	break;
	370	case -1:
	371	break;
	372	case 0:
	373	execl(cmd, cmd, event, dev, disc, NULL);
	374	exit(2);
	375	}
	376	}
	377	if (mailaddr && strncmp(event, "Fail", 4)==0) {
	378	FILE *mp = popen(Sendmail, "w");
	379	if (mp) {
	380	char hname[256];
	381	gethostname(hname, sizeof(hname));
	382	signal(SIGPIPE, SIG_IGN);
	383	fprintf(mp, "From: " Name " monitoring <root>\n");
	384	fprintf(mp, "To: %s\n", mailaddr);
	385	fprintf(mp, "Subject: %s event on %s:%s\n\n", event, dev, hname);
	386
	387	fprintf(mp, "This is an automatically generated mail message from " Name "\n");
	388	fprintf(mp, "running on %s\n\n", hname);
	389
	390	fprintf(mp, "A %s event had been detected on md device %s.\n\n", event, dev);
	391
	392	if (disc)
b83d95f3	393	fprintf(mp, "It could be related to componenet device %s.\n\n", disc);
52826846 NB	394
	395	fprintf(mp, "Faithfully yours, etc.\n");
	396	fclose(mp);
	397	}
	398
	399	}
	400	/* FIXME log the event to syslog maybe */
	401	}