xref: /freebsd/cddl/usr.sbin/zfsd/zfsd.cc (revision a90b9d0159070121c221b966469c3e36d912bf82)
1 /*-
2  * Copyright (c) 2011, 2012, 2013, 2014, 2015, 2016  Spectra Logic Corporation
3  * All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice, this list of conditions, and the following disclaimer,
10  *    without modification.
11  * 2. Redistributions in binary form must reproduce at minimum a disclaimer
12  *    substantially similar to the "NO WARRANTY" disclaimer below
13  *    ("Disclaimer") and any redistribution must be conditioned upon
14  *    including a substantially similar Disclaimer requirement for further
15  *    binary redistribution.
16  *
17  * NO WARRANTY
18  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
19  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
20  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR
21  * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
22  * HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
26  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
27  * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
28  * POSSIBILITY OF SUCH DAMAGES.
29  *
30  * Authors: Justin T. Gibbs     (Spectra Logic Corporation)
31  */
32 
33 /**
34  * \file zfsd.cc
35  *
36  * The ZFS daemon consumes kernel devdctl(4) event data via devd(8)'s
37  * unix domain socket in order to react to system changes that impact
38  * the function of ZFS storage pools.  The goal of this daemon is to
39  * provide similar functionality to the Solaris ZFS Diagnostic Engine
40  * (zfs-diagnosis), the Solaris ZFS fault handler (zfs-retire), and
41  * the Solaris ZFS vdev insertion agent (zfs-mod sysevent handler).
42  */
43 
44 #include <sys/cdefs.h>
45 #include <sys/byteorder.h>
46 #include <sys/param.h>
47 #include <sys/fs/zfs.h>
48 
49 #include <err.h>
50 #include <fcntl.h>
51 #include <libgeom.h>
52 #include <libutil.h>
53 #include <poll.h>
54 #include <syslog.h>
55 
56 #include <libzfs.h>
57 
58 #include <list>
59 #include <map>
60 #include <string>
61 
62 #include <devdctl/guid.h>
63 #include <devdctl/event.h>
64 #include <devdctl/event_factory.h>
65 #include <devdctl/exception.h>
66 #include <devdctl/consumer.h>
67 
68 #include "callout.h"
69 #include "vdev_iterator.h"
70 #include "zfsd_event.h"
71 #include "case_file.h"
72 #include "vdev.h"
73 #include "vdev_iterator.h"
74 #include "zfsd.h"
75 #include "zfsd_exception.h"
76 #include "zpool_list.h"
77 /*================================== Macros ==================================*/
78 #define NUM_ELEMENTS(x) (sizeof(x) / sizeof(*x))
79 
80 /*============================ Namespace Control =============================*/
81 using DevdCtl::Event;
82 using DevdCtl::EventFactory;
83 using DevdCtl::EventList;
84 
85 /*================================ Global Data ===============================*/
86 int              g_debug = 0;
87 libzfs_handle_t *g_zfsHandle;
88 
89 /*--------------------------------- ZfsDaemon --------------------------------*/
90 //- ZfsDaemon Static Private Data ----------------------------------------------
91 ZfsDaemon	    *ZfsDaemon::s_theZfsDaemon;
92 bool		     ZfsDaemon::s_logCaseFiles;
93 bool		     ZfsDaemon::s_terminateEventLoop;
94 char		     ZfsDaemon::s_pidFilePath[] = "/var/run/zfsd.pid";
95 pidfh		    *ZfsDaemon::s_pidFH;
96 int		     ZfsDaemon::s_signalPipeFD[2];
97 bool		     ZfsDaemon::s_systemRescanRequested(false);
98 EventFactory::Record ZfsDaemon::s_registryEntries[] =
99 {
100 	{ Event::NOTIFY, "GEOM",  &GeomEvent::Builder },
101 	{ Event::NOTIFY, "ZFS",   &ZfsEvent::Builder }
102 };
103 
104 //- ZfsDaemon Static Public Methods --------------------------------------------
105 ZfsDaemon &
106 ZfsDaemon::Get()
107 {
108 	return (*s_theZfsDaemon);
109 }
110 
111 void
112 ZfsDaemon::WakeEventLoop()
113 {
114 	write(s_signalPipeFD[1], "+", 1);
115 }
116 
117 void
118 ZfsDaemon::RequestSystemRescan()
119 {
120 	s_systemRescanRequested = true;
121 	ZfsDaemon::WakeEventLoop();
122 }
123 
124 void
125 ZfsDaemon::Run()
126 {
127 	ZfsDaemon daemon;
128 
129 	while (s_terminateEventLoop == false) {
130 
131 		try {
132 			daemon.DisconnectFromDevd();
133 
134 			if (daemon.ConnectToDevd() == false) {
135 				sleep(30);
136 				continue;
137 			}
138 
139 			daemon.DetectMissedEvents();
140 
141 			daemon.EventLoop();
142 
143 		} catch (const DevdCtl::Exception &exp) {
144 			exp.Log();
145 		}
146 	}
147 
148 	daemon.DisconnectFromDevd();
149 }
150 
151 //- ZfsDaemon Private Methods --------------------------------------------------
152 ZfsDaemon::ZfsDaemon()
153  : Consumer(/*defBuilder*/NULL, s_registryEntries,
154 	    NUM_ELEMENTS(s_registryEntries))
155 {
156 	if (s_theZfsDaemon != NULL)
157 		errx(1, "Multiple ZfsDaemon instances created. Exiting");
158 
159 	s_theZfsDaemon = this;
160 
161 	if (pipe(s_signalPipeFD) != 0)
162 		errx(1, "Unable to allocate signal pipe. Exiting");
163 
164 	if (fcntl(s_signalPipeFD[0], F_SETFL, O_NONBLOCK) == -1)
165 		errx(1, "Unable to set pipe as non-blocking. Exiting");
166 
167 	if (fcntl(s_signalPipeFD[1], F_SETFL, O_NONBLOCK) == -1)
168 		errx(1, "Unable to set pipe as non-blocking. Exiting");
169 
170 	signal(SIGHUP,  ZfsDaemon::RescanSignalHandler);
171 	signal(SIGINFO, ZfsDaemon::InfoSignalHandler);
172 	signal(SIGINT,  ZfsDaemon::QuitSignalHandler);
173 	signal(SIGTERM, ZfsDaemon::QuitSignalHandler);
174 	signal(SIGUSR1, ZfsDaemon::RescanSignalHandler);
175 
176 	g_zfsHandle = libzfs_init();
177 	if (g_zfsHandle == NULL)
178 		errx(1, "Unable to initialize ZFS library. Exiting");
179 
180 	Callout::Init();
181 	InitializeSyslog();
182 	OpenPIDFile();
183 
184 	if (g_debug == 0)
185 		daemon(0, 0);
186 
187 	UpdatePIDFile();
188 }
189 
190 ZfsDaemon::~ZfsDaemon()
191 {
192 	PurgeCaseFiles();
193 	ClosePIDFile();
194 }
195 
196 void
197 ZfsDaemon::PurgeCaseFiles()
198 {
199 	CaseFile::PurgeAll();
200 }
201 
202 bool
203 ZfsDaemon::VdevAddCaseFile(Vdev &vdev, void *cbArg)
204 {
205 	if (vdev.State() != VDEV_STATE_HEALTHY)
206 		CaseFile::Create(vdev);
207 
208 	return (/*break early*/false);
209 }
210 
211 void
212 ZfsDaemon::BuildCaseFiles()
213 {
214 	ZpoolList zpl;
215 	ZpoolList::iterator pool;
216 
217 	/* Add CaseFiles for vdevs with issues. */
218 	for (pool = zpl.begin(); pool != zpl.end(); pool++)
219 		VdevIterator(*pool).Each(VdevAddCaseFile, NULL);
220 
221 	/* De-serialize any saved cases. */
222 	CaseFile::DeSerialize();
223 
224 	/* Simulate config_sync events to force CaseFile reevaluation */
225 	for (pool = zpl.begin(); pool != zpl.end(); pool++) {
226 		char evString[160];
227 		Event *event;
228 		nvlist_t *config;
229 		uint64_t poolGUID;
230 		const char *poolname;
231 
232 		poolname = zpool_get_name(*pool);
233 		config = zpool_get_config(*pool, NULL);
234 		if (config == NULL) {
235 			syslog(LOG_ERR, "ZFSDaemon::BuildCaseFiles: Could not "
236 			    "find pool config for pool %s", poolname);
237 			continue;
238 		}
239 		if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID,
240 				     &poolGUID) != 0) {
241 			syslog(LOG_ERR, "ZFSDaemon::BuildCaseFiles: Could not "
242 			    "find pool guid for pool %s", poolname);
243 			continue;
244 		}
245 
246 
247 		snprintf(evString, 160, "!system=ZFS subsystem=ZFS "
248 		    "type=sysevent.fs.zfs.config_sync sub_type=synthesized "
249 		    "pool_name=%s pool_guid=%" PRIu64 "\n", poolname, poolGUID);
250 		event = Event::CreateEvent(GetFactory(), string(evString));
251 		if (event != NULL) {
252 			event->Process();
253 			delete event;
254 		}
255 	}
256 }
257 
258 void
259 ZfsDaemon::RescanSystem()
260 {
261         struct gmesh	  mesh;
262         struct gclass	 *mp;
263         struct ggeom	 *gp;
264         struct gprovider *pp;
265 	int		  result;
266 
267         /*
268 	 * The devdctl system doesn't replay events for new consumers
269 	 * of the interface.  Emit manufactured DEVFS arrival events
270 	 * for any devices that already before we started or during
271 	 * periods where we've lost our connection to devd.
272          */
273 	result = geom_gettree(&mesh);
274 	if (result != 0) {
275 		syslog(LOG_ERR, "ZfsDaemon::RescanSystem: "
276 		       "geom_gettree failed with error %d\n", result);
277 		return;
278 	}
279 
280 	const string evStart("!system=DEVFS subsystem=CDEV type=CREATE "
281 			     "sub_type=synthesized cdev=");
282         LIST_FOREACH(mp, &mesh.lg_class, lg_class) {
283                 LIST_FOREACH(gp, &mp->lg_geom, lg_geom) {
284                         LIST_FOREACH(pp, &gp->lg_provider, lg_provider) {
285 				Event *event;
286 
287 				string evString(evStart + pp->lg_name + "\n");
288 				event = Event::CreateEvent(GetFactory(),
289 							   evString);
290 				if (event != NULL) {
291 					if (event->Process())
292 						SaveEvent(*event);
293 					delete event;
294 				}
295                         }
296                 }
297 	}
298 	geom_deletetree(&mesh);
299 }
300 
301 void
302 ZfsDaemon::DetectMissedEvents()
303 {
304 	do {
305 		PurgeCaseFiles();
306 
307 		/*
308 		 * Discard any events waiting for us.  We don't know
309 		 * if they still apply to the current state of the
310 		 * system.
311 		 */
312 		FlushEvents();
313 
314 		BuildCaseFiles();
315 
316 		/*
317 		 * If the system state has changed during our
318 		 * interrogation, start over.
319 		 */
320 	} while (s_terminateEventLoop == false && EventsPending());
321 
322 	RescanSystem();
323 }
324 
325 void
326 ZfsDaemon::EventLoop()
327 {
328 	while (s_terminateEventLoop == false) {
329 		struct pollfd fds[2];
330 		int	      result;
331 
332 		if (s_logCaseFiles == true) {
333 			EventList::iterator event(m_unconsumedEvents.begin());
334 			s_logCaseFiles = false;
335 			CaseFile::LogAll();
336 			while (event != m_unconsumedEvents.end())
337 				(*event++)->Log(LOG_INFO);
338 		}
339 
340 		Callout::ExpireCallouts();
341 
342 		/* Wait for data. */
343 		fds[0].fd      = m_devdSockFD;
344 		fds[0].events  = POLLIN;
345 		fds[0].revents = 0;
346 		fds[1].fd      = s_signalPipeFD[0];
347 		fds[1].events  = POLLIN;
348 		fds[1].revents = 0;
349 		result = poll(fds, NUM_ELEMENTS(fds), /*timeout*/INFTIM);
350 		if (result == -1) {
351 			if (errno == EINTR)
352 				continue;
353 			else
354 				err(1, "Polling for devd events failed");
355 		} else if (result == 0) {
356 			errx(1, "Unexpected result of 0 from poll. Exiting");
357 		}
358 
359 		if ((fds[0].revents & POLLIN) != 0)
360 			ProcessEvents();
361 
362 		if ((fds[1].revents & POLLIN) != 0) {
363 			static char discardBuf[128];
364 
365 			/*
366 			 * This pipe exists just to close the signal
367 			 * race.  Its contents are of no interest to
368 			 * us, but we must ensure that future signals
369 			 * have space in the pipe to write.
370 			 */
371 			while (read(s_signalPipeFD[0], discardBuf,
372 				    sizeof(discardBuf)) > 0)
373 				;
374 		}
375 
376 		if (s_systemRescanRequested == true) {
377 			s_systemRescanRequested = false;
378 			syslog(LOG_INFO, "System Rescan request processed.");
379 			RescanSystem();
380 		}
381 
382 		if ((fds[0].revents & POLLERR) != 0) {
383 			syslog(LOG_INFO, "POLLERROR detected on devd socket.");
384 			break;
385 		}
386 
387 		if ((fds[0].revents & POLLHUP) != 0) {
388 			syslog(LOG_INFO, "POLLHUP detected on devd socket.");
389 			break;
390 		}
391 	}
392 }
393 //- ZfsDaemon staic Private Methods --------------------------------------------
394 void
395 ZfsDaemon::InfoSignalHandler(int)
396 {
397 	s_logCaseFiles = true;
398 	ZfsDaemon::WakeEventLoop();
399 }
400 
401 void
402 ZfsDaemon::RescanSignalHandler(int)
403 {
404 	RequestSystemRescan();
405 }
406 
407 void
408 ZfsDaemon::QuitSignalHandler(int)
409 {
410 	s_terminateEventLoop = true;
411 	ZfsDaemon::WakeEventLoop();
412 }
413 
414 void
415 ZfsDaemon::OpenPIDFile()
416 {
417 	pid_t otherPID;
418 
419 	s_pidFH = pidfile_open(s_pidFilePath, 0600, &otherPID);
420 	if (s_pidFH == NULL) {
421 		if (errno == EEXIST)
422 			errx(1, "already running as PID %d. Exiting", otherPID);
423 		warn("cannot open PID file");
424 	}
425 }
426 
427 void
428 ZfsDaemon::UpdatePIDFile()
429 {
430 	if (s_pidFH != NULL)
431 		pidfile_write(s_pidFH);
432 }
433 
434 void
435 ZfsDaemon::ClosePIDFile()
436 {
437 	if (s_pidFH != NULL)
438 		pidfile_remove(s_pidFH);
439 }
440 
441 void
442 ZfsDaemon::InitializeSyslog()
443 {
444 	openlog("zfsd", LOG_NDELAY, LOG_DAEMON);
445 }
446 
447