?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
ctdbd: remove a nonempty blank line
[Samba.git] / ctdb / server / ctdb_call.c
blobfeb3e61ee766ab58760f32ca70b90ef4042c11cf
1 /*
2 ctdb_call protocol code
3
4 Copyright (C) Andrew Tridgell 2006
5
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 3 of the License, or
9 (at your option) any later version.
10
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
15
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, see <http://www.gnu.org/licenses/>.
18 */
19 /*
20 see http://wiki.samba.org/index.php/Samba_%26_Clustering for
21 protocol design and packet details
22 */
23 #include "includes.h"
24 #include "lib/tdb/include/tdb.h"
25 #include "lib/util/dlinklist.h"
26 #include "system/network.h"
27 #include "system/filesys.h"
28 #include "../include/ctdb_private.h"
29 #include "../common/rb_tree.h"
30
31 struct ctdb_sticky_record {
32 struct ctdb_context *ctdb;
33 struct ctdb_db_context *ctdb_db;
34 TDB_CONTEXT *pindown;
35 };
36
37 /*
38 find the ctdb_db from a db index
39 */
40 struct ctdb_db_context *find_ctdb_db(struct ctdb_context *ctdb, uint32_t id)
41 {
42 struct ctdb_db_context *ctdb_db;
43
44 for (ctdb_db=ctdb->db_list; ctdb_db; ctdb_db=ctdb_db->next) {
45 if (ctdb_db->db_id == id) {
46 break;
47 }
48 }
49 return ctdb_db;
50 }
51
52 /*
53 a varient of input packet that can be used in lock requeue
54 */
55 static void ctdb_call_input_pkt(void *p, struct ctdb_req_header *hdr)
56 {
57 struct ctdb_context *ctdb = talloc_get_type(p, struct ctdb_context);
58 ctdb_input_pkt(ctdb, hdr);
59 }
60
61
62 /*
63 send an error reply
64 */
65 static void ctdb_send_error(struct ctdb_context *ctdb,
66 struct ctdb_req_header *hdr, uint32_t status,
67 const char *fmt, ...) PRINTF_ATTRIBUTE(4,5);
68 static void ctdb_send_error(struct ctdb_context *ctdb,
69 struct ctdb_req_header *hdr, uint32_t status,
70 const char *fmt, ...)
71 {
72 va_list ap;
73 struct ctdb_reply_error *r;
74 char *msg;
75 int msglen, len;
76
77 if (ctdb->methods == NULL) {
78 DEBUG(DEBUG_INFO,(__location__ " Failed to send error. Transport is DOWN\n"));
79 return;
80 }
81
82 va_start(ap, fmt);
83 msg = talloc_vasprintf(ctdb, fmt, ap);
84 if (msg == NULL) {
85 ctdb_fatal(ctdb, "Unable to allocate error in ctdb_send_error\n");
86 }
87 va_end(ap);
88
89 msglen = strlen(msg)+1;
90 len = offsetof(struct ctdb_reply_error, msg);
91 r = ctdb_transport_allocate(ctdb, msg, CTDB_REPLY_ERROR, len + msglen,
92 struct ctdb_reply_error);
93 CTDB_NO_MEMORY_FATAL(ctdb, r);
94
95 r->hdr.destnode = hdr->srcnode;
96 r->hdr.reqid = hdr->reqid;
97 r->status = status;
98 r->msglen = msglen;
99 memcpy(&r->msg[0], msg, msglen);
100
101 ctdb_queue_packet(ctdb, &r->hdr);
102
103 talloc_free(msg);
104 }
105
106
107 /**
108 * send a redirect reply
109 *
110 * The logic behind this function is this:
111 *
112 * A client wants to grab a record and sends a CTDB_REQ_CALL packet
113 * to its local ctdb (ctdb_request_call). If the node is not itself
114 * the record's DMASTER, it first redirects the packet to the
115 * record's LMASTER. The LMASTER then redirects the call packet to
116 * the current DMASTER. Note that this works because of this: When
117 * a record is migrated off a node, then the new DMASTER is stored
118 * in the record's copy on the former DMASTER.
119 */
120 static void ctdb_call_send_redirect(struct ctdb_context *ctdb,
121 struct ctdb_db_context *ctdb_db,
122 TDB_DATA key,
123 struct ctdb_req_call *c,
124 struct ctdb_ltdb_header *header)
125 {
126 uint32_t lmaster = ctdb_lmaster(ctdb, &key);
127
128 c->hdr.destnode = lmaster;
129 if (ctdb->pnn == lmaster) {
130 c->hdr.destnode = header->dmaster;
131 }
132 c->hopcount++;
133
134 if (c->hopcount%100 == 99) {
135 DEBUG(DEBUG_WARNING,("High hopcount %d dbid:0x%08x "
136 "key:0x%08x pnn:%d src:%d lmaster:%d "
137 "header->dmaster:%d dst:%d\n",
138 c->hopcount, ctdb_db->db_id, ctdb_hash(&key),
139 ctdb->pnn, c->hdr.srcnode, lmaster,
140 header->dmaster, c->hdr.destnode));
141 }
142
143 ctdb_queue_packet(ctdb, &c->hdr);
144 }
145
146
147 /*
148 send a dmaster reply
149
150 caller must have the chainlock before calling this routine. Caller must be
151 the lmaster
152 */
153 static void ctdb_send_dmaster_reply(struct ctdb_db_context *ctdb_db,
154 struct ctdb_ltdb_header *header,
155 TDB_DATA key, TDB_DATA data,
156 uint32_t new_dmaster,
157 uint32_t reqid)
158 {
159 struct ctdb_context *ctdb = ctdb_db->ctdb;
160 struct ctdb_reply_dmaster *r;
161 int ret, len;
162 TALLOC_CTX *tmp_ctx;
163
164 if (ctdb->pnn != ctdb_lmaster(ctdb, &key)) {
165 DEBUG(DEBUG_ALERT,(__location__ " Caller is not lmaster!\n"));
166 return;
167 }
168
169 header->dmaster = new_dmaster;
170 ret = ctdb_ltdb_store(ctdb_db, key, header, data);
171 if (ret != 0) {
172 ctdb_fatal(ctdb, "ctdb_send_dmaster_reply unable to update dmaster");
173 return;
174 }
175
176 if (ctdb->methods == NULL) {
177 ctdb_fatal(ctdb, "ctdb_send_dmaster_reply cant update dmaster since transport is down");
178 return;
179 }
180
181 /* put the packet on a temporary context, allowing us to safely free
182 it below even if ctdb_reply_dmaster() has freed it already */
183 tmp_ctx = talloc_new(ctdb);
184
185 /* send the CTDB_REPLY_DMASTER */
186 len = offsetof(struct ctdb_reply_dmaster, data) + key.dsize + data.dsize + sizeof(uint32_t);
187 r = ctdb_transport_allocate(ctdb, tmp_ctx, CTDB_REPLY_DMASTER, len,
188 struct ctdb_reply_dmaster);
189 CTDB_NO_MEMORY_FATAL(ctdb, r);
190
191 r->hdr.destnode = new_dmaster;
192 r->hdr.reqid = reqid;
193 r->rsn = header->rsn;
194 r->keylen = key.dsize;
195 r->datalen = data.dsize;
196 r->db_id = ctdb_db->db_id;
197 memcpy(&r->data[0], key.dptr, key.dsize);
198 memcpy(&r->data[key.dsize], data.dptr, data.dsize);
199 memcpy(&r->data[key.dsize+data.dsize], &header->flags, sizeof(uint32_t));
200
201 ctdb_queue_packet(ctdb, &r->hdr);
202
203 talloc_free(tmp_ctx);
204 }
205
206 /*
207 send a dmaster request (give another node the dmaster for a record)
208
209 This is always sent to the lmaster, which ensures that the lmaster
210 always knows who the dmaster is. The lmaster will then send a
211 CTDB_REPLY_DMASTER to the new dmaster
212 */
213 static void ctdb_call_send_dmaster(struct ctdb_db_context *ctdb_db,
214 struct ctdb_req_call *c,
215 struct ctdb_ltdb_header *header,
216 TDB_DATA *key, TDB_DATA *data)
217 {
218 struct ctdb_req_dmaster *r;
219 struct ctdb_context *ctdb = ctdb_db->ctdb;
220 int len;
221 uint32_t lmaster = ctdb_lmaster(ctdb, key);
222
223 if (ctdb->methods == NULL) {
224 ctdb_fatal(ctdb, "Failed ctdb_call_send_dmaster since transport is down");
225 return;
226 }
227
228 if (data->dsize != 0) {
229 header->flags |= CTDB_REC_FLAG_MIGRATED_WITH_DATA;
230 }
231
232 if (lmaster == ctdb->pnn) {
233 ctdb_send_dmaster_reply(ctdb_db, header, *key, *data,
234 c->hdr.srcnode, c->hdr.reqid);
235 return;
236 }
237
238 len = offsetof(struct ctdb_req_dmaster, data) + key->dsize + data->dsize
239 + sizeof(uint32_t);
240 r = ctdb_transport_allocate(ctdb, ctdb, CTDB_REQ_DMASTER, len,
241 struct ctdb_req_dmaster);
242 CTDB_NO_MEMORY_FATAL(ctdb, r);
243 r->hdr.destnode = lmaster;
244 r->hdr.reqid = c->hdr.reqid;
245 r->db_id = c->db_id;
246 r->rsn = header->rsn;
247 r->dmaster = c->hdr.srcnode;
248 r->keylen = key->dsize;
249 r->datalen = data->dsize;
250 memcpy(&r->data[0], key->dptr, key->dsize);
251 memcpy(&r->data[key->dsize], data->dptr, data->dsize);
252 memcpy(&r->data[key->dsize + data->dsize], &header->flags, sizeof(uint32_t));
253
254 header->dmaster = c->hdr.srcnode;
255 if (ctdb_ltdb_store(ctdb_db, *key, header, *data) != 0) {
256 ctdb_fatal(ctdb, "Failed to store record in ctdb_call_send_dmaster");
257 }
258
259 ctdb_queue_packet(ctdb, &r->hdr);
260
261 talloc_free(r);
262 }
263
264 static void ctdb_sticky_pindown_timeout(struct event_context *ev, struct timed_event *te,
265 struct timeval t, void *private_data)
266 {
267 struct ctdb_sticky_record *sr = talloc_get_type(private_data,
268 struct ctdb_sticky_record);
269
270 DEBUG(DEBUG_ERR,("Pindown timeout db:%s unstick record\n", sr->ctdb_db->db_name));
271 if (sr->pindown != NULL) {
272 talloc_free(sr->pindown);
273 sr->pindown = NULL;
274 }
275 }
276
277 static int
278 ctdb_set_sticky_pindown(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, TDB_DATA key)
279 {
280 TALLOC_CTX *tmp_ctx = talloc_new(NULL);
281 uint32_t *k;
282 struct ctdb_sticky_record *sr;
283
284 k = talloc_zero_size(tmp_ctx, ((key.dsize + 3) & 0xfffffffc) + 4);
285 if (k == NULL) {
286 DEBUG(DEBUG_ERR,("Failed to allocate key for sticky record\n"));
287 talloc_free(tmp_ctx);
288 return -1;
289 }
290
291 k[0] = (key.dsize + 3) / 4 + 1;
292 memcpy(&k[1], key.dptr, key.dsize);
293
294 sr = trbt_lookuparray32(ctdb_db->sticky_records, k[0], &k[0]);
295 if (sr == NULL) {
296 talloc_free(tmp_ctx);
297 return 0;
298 }
299
300 talloc_free(tmp_ctx);
301
302 if (sr->pindown == NULL) {
303 DEBUG(DEBUG_ERR,("Pinning down record in %s for %d ms\n", ctdb_db->db_name, ctdb->tunable.sticky_pindown));
304 sr->pindown = talloc_new(sr);
305 if (sr->pindown == NULL) {
306 DEBUG(DEBUG_ERR,("Failed to allocate pindown context for sticky record\n"));
307 return -1;
308 }
309 event_add_timed(ctdb->ev, sr->pindown, timeval_current_ofs(ctdb->tunable.sticky_pindown / 1000, (ctdb->tunable.sticky_pindown * 1000) % 1000000), ctdb_sticky_pindown_timeout, sr);
310 }
311
312 return 0;
313 }
314
315 /*
316 called when a CTDB_REPLY_DMASTER packet comes in, or when the lmaster
317 gets a CTDB_REQUEST_DMASTER for itself. We become the dmaster.
318
319 must be called with the chainlock held. This function releases the chainlock
320 */
321 static void ctdb_become_dmaster(struct ctdb_db_context *ctdb_db,
322 struct ctdb_req_header *hdr,
323 TDB_DATA key, TDB_DATA data,
324 uint64_t rsn, uint32_t record_flags)
325 {
326 struct ctdb_call_state *state;
327 struct ctdb_context *ctdb = ctdb_db->ctdb;
328 struct ctdb_ltdb_header header;
329 int ret;
330
331 DEBUG(DEBUG_DEBUG,("pnn %u dmaster response %08x\n", ctdb->pnn, ctdb_hash(&key)));
332
333 ZERO_STRUCT(header);
334 header.rsn = rsn;
335 header.dmaster = ctdb->pnn;
336 header.flags = record_flags;
337
338 state = ctdb_reqid_find(ctdb, hdr->reqid, struct ctdb_call_state);
339
340 if (state) {
341 if (state->call->flags & CTDB_CALL_FLAG_VACUUM_MIGRATION) {
342 /*
343 * We temporarily add the VACUUM_MIGRATED flag to
344 * the record flags, so that ctdb_ltdb_store can
345 * decide whether the record should be stored or
346 * deleted.
347 */
348 header.flags |= CTDB_REC_FLAG_VACUUM_MIGRATED;
349 }
350 }
351
352 if (ctdb_ltdb_store(ctdb_db, key, &header, data) != 0) {
353 ctdb_fatal(ctdb, "ctdb_reply_dmaster store failed\n");
354
355 ret = ctdb_ltdb_unlock(ctdb_db, key);
356 if (ret != 0) {
357 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
358 }
359 return;
360 }
361
362 /* we just became DMASTER and this database is "sticky",
363 see if the record is flagged as "hot" and set up a pin-down
364 context to stop migrations for a little while if so
365 */
366 if (ctdb_db->sticky) {
367 ctdb_set_sticky_pindown(ctdb, ctdb_db, key);
368 }
369
370 if (state == NULL) {
371 DEBUG(DEBUG_ERR,("pnn %u Invalid reqid %u in ctdb_become_dmaster from node %u\n",
372 ctdb->pnn, hdr->reqid, hdr->srcnode));
373
374 ret = ctdb_ltdb_unlock(ctdb_db, key);
375 if (ret != 0) {
376 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
377 }
378 return;
379 }
380
381 if (key.dsize != state->call->key.dsize || memcmp(key.dptr, state->call->key.dptr, key.dsize)) {
382 DEBUG(DEBUG_ERR, ("Got bogus DMASTER packet reqid:%u from node %u. Key does not match key held in matching idr.\n", hdr->reqid, hdr->srcnode));
383
384 ret = ctdb_ltdb_unlock(ctdb_db, key);
385 if (ret != 0) {
386 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
387 }
388 return;
389 }
390
391 if (hdr->reqid != state->reqid) {
392 /* we found a record but it was the wrong one */
393 DEBUG(DEBUG_ERR, ("Dropped orphan in ctdb_become_dmaster with reqid:%u\n from node %u", hdr->reqid, hdr->srcnode));
394
395 ret = ctdb_ltdb_unlock(ctdb_db, key);
396 if (ret != 0) {
397 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
398 }
399 return;
400 }
401
402 ctdb_call_local(ctdb_db, state->call, &header, state, &data, true, ctdb->pnn);
403
404 ret = ctdb_ltdb_unlock(ctdb_db, state->call->key);
405 if (ret != 0) {
406 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
407 }
408
409 state->state = CTDB_CALL_DONE;
410 if (state->async.fn) {
411 state->async.fn(state);
412 }
413 }
414
415
416
417 /*
418 called when a CTDB_REQ_DMASTER packet comes in
419
420 this comes into the lmaster for a record when the current dmaster
421 wants to give up the dmaster role and give it to someone else
422 */
423 void ctdb_request_dmaster(struct ctdb_context *ctdb, struct ctdb_req_header *hdr)
424 {
425 struct ctdb_req_dmaster *c = (struct ctdb_req_dmaster *)hdr;
426 TDB_DATA key, data, data2;
427 struct ctdb_ltdb_header header;
428 struct ctdb_db_context *ctdb_db;
429 uint32_t record_flags = 0;
430 size_t len;
431 int ret;
432
433 key.dptr = c->data;
434 key.dsize = c->keylen;
435 data.dptr = c->data + c->keylen;
436 data.dsize = c->datalen;
437 len = offsetof(struct ctdb_req_dmaster, data) + key.dsize + data.dsize
438 + sizeof(uint32_t);
439 if (len <= c->hdr.length) {
440 record_flags = *(uint32_t *)&c->data[c->keylen + c->datalen];
441 }
442
443 ctdb_db = find_ctdb_db(ctdb, c->db_id);
444 if (!ctdb_db) {
445 ctdb_send_error(ctdb, hdr, -1,
446 "Unknown database in request. db_id==0x%08x",
447 c->db_id);
448 return;
449 }
450
451 /* fetch the current record */
452 ret = ctdb_ltdb_lock_fetch_requeue(ctdb_db, key, &header, hdr, &data2,
453 ctdb_call_input_pkt, ctdb, false);
454 if (ret == -1) {
455 ctdb_fatal(ctdb, "ctdb_req_dmaster failed to fetch record");
456 return;
457 }
458 if (ret == -2) {
459 DEBUG(DEBUG_INFO,(__location__ " deferring ctdb_request_dmaster\n"));
460 return;
461 }
462
463 if (ctdb_lmaster(ctdb, &key) != ctdb->pnn) {
464 DEBUG(DEBUG_ALERT,("pnn %u dmaster request to non-lmaster lmaster=%u gen=%u curgen=%u\n",
465 ctdb->pnn, ctdb_lmaster(ctdb, &key),
466 hdr->generation, ctdb->vnn_map->generation));
467 ctdb_fatal(ctdb, "ctdb_req_dmaster to non-lmaster");
468 }
469
470 DEBUG(DEBUG_DEBUG,("pnn %u dmaster request on %08x for %u from %u\n",
471 ctdb->pnn, ctdb_hash(&key), c->dmaster, c->hdr.srcnode));
472
473 /* its a protocol error if the sending node is not the current dmaster */
474 if (header.dmaster != hdr->srcnode) {
475 DEBUG(DEBUG_ALERT,("pnn %u dmaster request for new-dmaster %u from non-master %u real-dmaster=%u key %08x dbid 0x%08x gen=%u curgen=%u c->rsn=%llu header.rsn=%llu reqid=%u keyval=0x%08x\n",
476 ctdb->pnn, c->dmaster, hdr->srcnode, header.dmaster, ctdb_hash(&key),
477 ctdb_db->db_id, hdr->generation, ctdb->vnn_map->generation,
478 (unsigned long long)c->rsn, (unsigned long long)header.rsn, c->hdr.reqid,
479 (key.dsize >= 4)?(*(uint32_t *)key.dptr):0));
480 if (header.rsn != 0 || header.dmaster != ctdb->pnn) {
481 DEBUG(DEBUG_ERR,("ctdb_req_dmaster from non-master. Force a recovery.\n"));
482
483 ctdb->recovery_mode = CTDB_RECOVERY_ACTIVE;
484 ctdb_ltdb_unlock(ctdb_db, key);
485 return;
486 }
487 }
488
489 if (header.rsn > c->rsn) {
490 DEBUG(DEBUG_ALERT,("pnn %u dmaster request with older RSN new-dmaster %u from %u real-dmaster=%u key %08x dbid 0x%08x gen=%u curgen=%u c->rsn=%llu header.rsn=%llu reqid=%u\n",
491 ctdb->pnn, c->dmaster, hdr->srcnode, header.dmaster, ctdb_hash(&key),
492 ctdb_db->db_id, hdr->generation, ctdb->vnn_map->generation,
493 (unsigned long long)c->rsn, (unsigned long long)header.rsn, c->hdr.reqid));
494 }
495
496 /* use the rsn from the sending node */
497 header.rsn = c->rsn;
498
499 /* store the record flags from the sending node */
500 header.flags = record_flags;
501
502 /* check if the new dmaster is the lmaster, in which case we
503 skip the dmaster reply */
504 if (c->dmaster == ctdb->pnn) {
505 ctdb_become_dmaster(ctdb_db, hdr, key, data, c->rsn, record_flags);
506 } else {
507 ctdb_send_dmaster_reply(ctdb_db, &header, key, data, c->dmaster, hdr->reqid);
508
509 ret = ctdb_ltdb_unlock(ctdb_db, key);
510 if (ret != 0) {
511 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
512 }
513 }
514 }
515
516 static void ctdb_sticky_record_timeout(struct event_context *ev, struct timed_event *te,
517 struct timeval t, void *private_data)
518 {
519 struct ctdb_sticky_record *sr = talloc_get_type(private_data,
520 struct ctdb_sticky_record);
521 talloc_free(sr);
522 }
523
524 static void *ctdb_make_sticky_record_callback(void *parm, void *data)
525 {
526 if (data) {
527 DEBUG(DEBUG_ERR,("Already have sticky record registered. Free old %p and create new %p\n", data, parm));
528 talloc_free(data);
529 }
530 return parm;
531 }
532
533 static int
534 ctdb_make_record_sticky(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, TDB_DATA key)
535 {
536 TALLOC_CTX *tmp_ctx = talloc_new(NULL);
537 uint32_t *k;
538 struct ctdb_sticky_record *sr;
539
540 k = talloc_zero_size(tmp_ctx, ((key.dsize + 3) & 0xfffffffc) + 4);
541 if (k == NULL) {
542 DEBUG(DEBUG_ERR,("Failed to allocate key for sticky record\n"));
543 talloc_free(tmp_ctx);
544 return -1;
545 }
546
547 k[0] = (key.dsize + 3) / 4 + 1;
548 memcpy(&k[1], key.dptr, key.dsize);
549
550 sr = trbt_lookuparray32(ctdb_db->sticky_records, k[0], &k[0]);
551 if (sr != NULL) {
552 talloc_free(tmp_ctx);
553 return 0;
554 }
555
556 sr = talloc(ctdb_db->sticky_records, struct ctdb_sticky_record);
557 if (sr == NULL) {
558 talloc_free(tmp_ctx);
559 DEBUG(DEBUG_ERR,("Failed to allocate sticky record structure\n"));
560 return -1;
561 }
562
563 sr->ctdb = ctdb;
564 sr->ctdb_db = ctdb_db;
565 sr->pindown = NULL;
566
567 DEBUG(DEBUG_ERR,("Make record sticky in db %s\n", ctdb_db->db_name));
568
569 trbt_insertarray32_callback(ctdb_db->sticky_records, k[0], &k[0], ctdb_make_sticky_record_callback, sr);
570
571 event_add_timed(ctdb->ev, sr, timeval_current_ofs(ctdb->tunable.sticky_duration, 0), ctdb_sticky_record_timeout, sr);
572
573 talloc_free(tmp_ctx);
574 return 0;
575 }
576
577 struct pinned_down_requeue_handle {
578 struct ctdb_context *ctdb;
579 struct ctdb_req_header *hdr;
580 };
581
582 struct pinned_down_deferred_call {
583 struct ctdb_context *ctdb;
584 struct ctdb_req_header *hdr;
585 };
586
587 static void pinned_down_requeue(struct event_context *ev, struct timed_event *te,
588 struct timeval t, void *private_data)
589 {
590 struct pinned_down_requeue_handle *handle = talloc_get_type(private_data, struct pinned_down_requeue_handle);
591 struct ctdb_context *ctdb = handle->ctdb;
592
593 talloc_steal(ctdb, handle->hdr);
594 ctdb_call_input_pkt(ctdb, handle->hdr);
595
596 talloc_free(handle);
597 }
598
599 static int pinned_down_destructor(struct pinned_down_deferred_call *pinned_down)
600 {
601 struct ctdb_context *ctdb = pinned_down->ctdb;
602 struct pinned_down_requeue_handle *handle = talloc(ctdb, struct pinned_down_requeue_handle);
603
604 handle->ctdb = pinned_down->ctdb;
605 handle->hdr = pinned_down->hdr;
606 talloc_steal(handle, handle->hdr);
607
608 event_add_timed(ctdb->ev, handle, timeval_zero(), pinned_down_requeue, handle);
609
610 return 0;
611 }
612
613 static int
614 ctdb_defer_pinned_down_request(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, TDB_DATA key, struct ctdb_req_header *hdr)
615 {
616 TALLOC_CTX *tmp_ctx = talloc_new(NULL);
617 uint32_t *k;
618 struct ctdb_sticky_record *sr;
619 struct pinned_down_deferred_call *pinned_down;
620
621 k = talloc_zero_size(tmp_ctx, ((key.dsize + 3) & 0xfffffffc) + 4);
622 if (k == NULL) {
623 DEBUG(DEBUG_ERR,("Failed to allocate key for sticky record\n"));
624 talloc_free(tmp_ctx);
625 return -1;
626 }
627
628 k[0] = (key.dsize + 3) / 4 + 1;
629 memcpy(&k[1], key.dptr, key.dsize);
630
631 sr = trbt_lookuparray32(ctdb_db->sticky_records, k[0], &k[0]);
632 if (sr == NULL) {
633 talloc_free(tmp_ctx);
634 return -1;
635 }
636
637 talloc_free(tmp_ctx);
638
639 if (sr->pindown == NULL) {
640 return -1;
641 }
642
643 pinned_down = talloc(sr->pindown, struct pinned_down_deferred_call);
644 if (pinned_down == NULL) {
645 DEBUG(DEBUG_ERR,("Failed to allocate structure for deferred pinned down request\n"));
646 return -1;
647 }
648
649 pinned_down->ctdb = ctdb;
650 pinned_down->hdr = hdr;
651
652 talloc_set_destructor(pinned_down, pinned_down_destructor);
653 talloc_steal(pinned_down, hdr);
654
655 return 0;
656 }
657
658 static void
659 ctdb_update_db_stat_hot_keys(struct ctdb_db_context *ctdb_db, TDB_DATA key, int hopcount)
660 {
661 int i;
662
663 /* smallest value is always at index 0 */
664 if (hopcount <= ctdb_db->statistics.hot_keys[0].count) {
665 return;
666 }
667
668 /* see if we already know this key */
669 for (i = 0; i < MAX_HOT_KEYS; i++) {
670 if (key.dsize != ctdb_db->statistics.hot_keys[i].key.dsize) {
671 continue;
672 }
673 if (memcmp(key.dptr, ctdb_db->statistics.hot_keys[i].key.dptr, key.dsize)) {
674 continue;
675 }
676 /* found an entry for this key */
677 if (hopcount <= ctdb_db->statistics.hot_keys[i].count) {
678 return;
679 }
680 ctdb_db->statistics.hot_keys[i].count = hopcount;
681 goto sort_keys;
682 }
683
684 if (ctdb_db->statistics.hot_keys[0].key.dptr != NULL) {
685 talloc_free(ctdb_db->statistics.hot_keys[0].key.dptr);
686 }
687 ctdb_db->statistics.hot_keys[0].key.dsize = key.dsize;
688 ctdb_db->statistics.hot_keys[0].key.dptr = talloc_memdup(ctdb_db, key.dptr, key.dsize);
689 ctdb_db->statistics.hot_keys[0].count = hopcount;
690
691
692 sort_keys:
693 for (i = 2; i < MAX_HOT_KEYS; i++) {
694 if (ctdb_db->statistics.hot_keys[i].count < ctdb_db->statistics.hot_keys[0].count) {
695 hopcount = ctdb_db->statistics.hot_keys[i].count;
696 ctdb_db->statistics.hot_keys[i].count = ctdb_db->statistics.hot_keys[0].count;
697 ctdb_db->statistics.hot_keys[0].count = hopcount;
698
699 key = ctdb_db->statistics.hot_keys[i].key;
700 ctdb_db->statistics.hot_keys[i].key = ctdb_db->statistics.hot_keys[0].key;
701 ctdb_db->statistics.hot_keys[0].key = key;
702 }
703 }
704 }
705
706 /*
707 called when a CTDB_REQ_CALL packet comes in
708 */
709 void ctdb_request_call(struct ctdb_context *ctdb, struct ctdb_req_header *hdr)
710 {
711 struct ctdb_req_call *c = (struct ctdb_req_call *)hdr;
712 TDB_DATA data;
713 struct ctdb_reply_call *r;
714 int ret, len;
715 struct ctdb_ltdb_header header;
716 struct ctdb_call *call;
717 struct ctdb_db_context *ctdb_db;
718 int tmp_count, bucket;
719
720 if (ctdb->methods == NULL) {
721 DEBUG(DEBUG_INFO,(__location__ " Failed ctdb_request_call. Transport is DOWN\n"));
722 return;
723 }
724
725
726 ctdb_db = find_ctdb_db(ctdb, c->db_id);
727 if (!ctdb_db) {
728 ctdb_send_error(ctdb, hdr, -1,
729 "Unknown database in request. db_id==0x%08x",
730 c->db_id);
731 return;
732 }
733
734 call = talloc(hdr, struct ctdb_call);
735 CTDB_NO_MEMORY_FATAL(ctdb, call);
736
737 call->call_id = c->callid;
738 call->key.dptr = c->data;
739 call->key.dsize = c->keylen;
740 call->call_data.dptr = c->data + c->keylen;
741 call->call_data.dsize = c->calldatalen;
742 call->reply_data.dptr = NULL;
743 call->reply_data.dsize = 0;
744
745
746 /* If this record is pinned down we should defer the
747 request until the pindown times out
748 */
749 if (ctdb_db->sticky) {
750 if (ctdb_defer_pinned_down_request(ctdb, ctdb_db, call->key, hdr) == 0) {
751 DEBUG(DEBUG_WARNING,("Defer request for pinned down record in %s\n", ctdb_db->db_name));
752 return;
753 }
754 }
755
756
757 /* determine if we are the dmaster for this key. This also
758 fetches the record data (if any), thus avoiding a 2nd fetch of the data
759 if the call will be answered locally */
760
761 ret = ctdb_ltdb_lock_fetch_requeue(ctdb_db, call->key, &header, hdr, &data,
762 ctdb_call_input_pkt, ctdb, false);
763 if (ret == -1) {
764 ctdb_send_error(ctdb, hdr, ret, "ltdb fetch failed in ctdb_request_call");
765 return;
766 }
767 if (ret == -2) {
768 DEBUG(DEBUG_INFO,(__location__ " deferred ctdb_request_call\n"));
769 return;
770 }
771
772 /* Dont do READONLY if we dont have a tracking database */
773 if ((c->flags & CTDB_WANT_READONLY) && !ctdb_db->readonly) {
774 c->flags &= ~CTDB_WANT_READONLY;
775 }
776
777 if (header.flags & CTDB_REC_RO_REVOKE_COMPLETE) {
778 header.flags &= ~CTDB_REC_RO_FLAGS;
779 CTDB_INCREMENT_STAT(ctdb, total_ro_revokes);
780 CTDB_INCREMENT_DB_STAT(ctdb_db, db_ro_revokes);
781 if (ctdb_ltdb_store(ctdb_db, call->key, &header, data) != 0) {
782 ctdb_fatal(ctdb, "Failed to write header with cleared REVOKE flag");
783 }
784 /* and clear out the tracking data */
785 if (tdb_delete(ctdb_db->rottdb, call->key) != 0) {
786 DEBUG(DEBUG_ERR,(__location__ " Failed to clear out trackingdb record\n"));
787 }
788 }
789
790 /* if we are revoking, we must defer all other calls until the revoke
791 * had completed.
792 */
793 if (header.flags & CTDB_REC_RO_REVOKING_READONLY) {
794 talloc_free(data.dptr);
795 ret = ctdb_ltdb_unlock(ctdb_db, call->key);
796
797 if (ctdb_add_revoke_deferred_call(ctdb, ctdb_db, call->key, hdr, ctdb_call_input_pkt, ctdb) != 0) {
798 ctdb_fatal(ctdb, "Failed to add deferred call for revoke child");
799 }
800 talloc_free(call);
801 return;
802 }
803
804 /* if we are not the dmaster and are not hosting any delegations,
805 then send a redirect to the requesting node */
806 if ((header.dmaster != ctdb->pnn)
807 && (!(header.flags & CTDB_REC_RO_HAVE_DELEGATIONS)) ) {
808 talloc_free(data.dptr);
809 ctdb_call_send_redirect(ctdb, ctdb_db, call->key, c, &header);
810
811 ret = ctdb_ltdb_unlock(ctdb_db, call->key);
812 if (ret != 0) {
813 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
814 }
815 return;
816 }
817
818 if ( (!(c->flags & CTDB_WANT_READONLY))
819 && (header.flags & (CTDB_REC_RO_HAVE_DELEGATIONS|CTDB_REC_RO_HAVE_READONLY)) ) {
820 header.flags |= CTDB_REC_RO_REVOKING_READONLY;
821 if (ctdb_ltdb_store(ctdb_db, call->key, &header, data) != 0) {
822 ctdb_fatal(ctdb, "Failed to store record with HAVE_DELEGATIONS set");
823 }
824 ret = ctdb_ltdb_unlock(ctdb_db, call->key);
825
826 if (ctdb_start_revoke_ro_record(ctdb, ctdb_db, call->key, &header, data) != 0) {
827 ctdb_fatal(ctdb, "Failed to start record revoke");
828 }
829 talloc_free(data.dptr);
830
831 if (ctdb_add_revoke_deferred_call(ctdb, ctdb_db, call->key, hdr, ctdb_call_input_pkt, ctdb) != 0) {
832 ctdb_fatal(ctdb, "Failed to add deferred call for revoke child");
833 }
834 talloc_free(call);
835
836 return;
837 }
838
839 /* If this is the first request for delegation. bump rsn and set
840 * the delegations flag
841 */
842 if ((c->flags & CTDB_WANT_READONLY)
843 && (c->callid == CTDB_FETCH_WITH_HEADER_FUNC)
844 && (!(header.flags & CTDB_REC_RO_HAVE_DELEGATIONS))) {
845 header.rsn += 3;
846 header.flags |= CTDB_REC_RO_HAVE_DELEGATIONS;
847 if (ctdb_ltdb_store(ctdb_db, call->key, &header, data) != 0) {
848 ctdb_fatal(ctdb, "Failed to store record with HAVE_DELEGATIONS set");
849 }
850 }
851 if ((c->flags & CTDB_WANT_READONLY)
852 && (call->call_id == CTDB_FETCH_WITH_HEADER_FUNC)) {
853 TDB_DATA tdata;
854
855 tdata = tdb_fetch(ctdb_db->rottdb, call->key);
856 if (ctdb_trackingdb_add_pnn(ctdb, &tdata, c->hdr.srcnode) != 0) {
857 ctdb_fatal(ctdb, "Failed to add node to trackingdb");
858 }
859 if (tdb_store(ctdb_db->rottdb, call->key, tdata, TDB_REPLACE) != 0) {
860 ctdb_fatal(ctdb, "Failed to store trackingdb data");
861 }
862 free(tdata.dptr);
863
864 ret = ctdb_ltdb_unlock(ctdb_db, call->key);
865 if (ret != 0) {
866 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
867 }
868
869 len = offsetof(struct ctdb_reply_call, data) + data.dsize + sizeof(struct ctdb_ltdb_header);
870 r = ctdb_transport_allocate(ctdb, ctdb, CTDB_REPLY_CALL, len,
871 struct ctdb_reply_call);
872 CTDB_NO_MEMORY_FATAL(ctdb, r);
873 r->hdr.destnode = c->hdr.srcnode;
874 r->hdr.reqid = c->hdr.reqid;
875 r->status = 0;
876 r->datalen = data.dsize + sizeof(struct ctdb_ltdb_header);
877 header.rsn -= 2;
878 header.flags |= CTDB_REC_RO_HAVE_READONLY;
879 header.flags &= ~CTDB_REC_RO_HAVE_DELEGATIONS;
880 memcpy(&r->data[0], &header, sizeof(struct ctdb_ltdb_header));
881
882 if (data.dsize) {
883 memcpy(&r->data[sizeof(struct ctdb_ltdb_header)], data.dptr, data.dsize);
884 }
885
886 ctdb_queue_packet(ctdb, &r->hdr);
887 CTDB_INCREMENT_STAT(ctdb, total_ro_delegations);
888 CTDB_INCREMENT_DB_STAT(ctdb_db, db_ro_delegations);
889
890 talloc_free(r);
891 return;
892 }
893
894 CTDB_UPDATE_STAT(ctdb, max_hop_count, c->hopcount);
895 tmp_count = c->hopcount;
896 bucket = 0;
897 while (tmp_count) {
898 tmp_count >>= 2;
899 bucket++;
900 }
901 if (bucket >= MAX_COUNT_BUCKETS) {
902 bucket = MAX_COUNT_BUCKETS - 1;
903 }
904 CTDB_INCREMENT_STAT(ctdb, hop_count_bucket[bucket]);
905 CTDB_INCREMENT_DB_STAT(ctdb_db, hop_count_bucket[bucket]);
906 ctdb_update_db_stat_hot_keys(ctdb_db, call->key, c->hopcount);
907
908 /* If this database supports sticky records, then check if the
909 hopcount is big. If it is it means the record is hot and we
910 should make it sticky.
911 */
912 if (ctdb_db->sticky && c->hopcount >= ctdb->tunable.hopcount_make_sticky) {
913 DEBUG(DEBUG_ERR, ("Hot record in database %s. Hopcount is %d. Make record sticky for %d seconds\n", ctdb_db->db_name, c->hopcount, ctdb->tunable.sticky_duration));
914 ctdb_make_record_sticky(ctdb, ctdb_db, call->key);
915 }
916
917
918 /* if this nodes has done enough consecutive calls on the same record
919 then give them the record
920 or if the node requested an immediate migration
921 */
922 if ( c->hdr.srcnode != ctdb->pnn &&
923 ((header.laccessor == c->hdr.srcnode
924 && header.lacount >= ctdb->tunable.max_lacount
925 && ctdb->tunable.max_lacount != 0)
926 || (c->flags & CTDB_IMMEDIATE_MIGRATION)) ) {
927 if (ctdb_db->transaction_active) {
928 DEBUG(DEBUG_INFO, (__location__ " refusing migration"
929 " of key %s while transaction is active\n",
930 (char *)call->key.dptr));
931 } else {
932 DEBUG(DEBUG_DEBUG,("pnn %u starting migration of %08x to %u\n",
933 ctdb->pnn, ctdb_hash(&(call->key)), c->hdr.srcnode));
934 ctdb_call_send_dmaster(ctdb_db, c, &header, &(call->key), &data);
935 talloc_free(data.dptr);
936
937 ret = ctdb_ltdb_unlock(ctdb_db, call->key);
938 if (ret != 0) {
939 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
940 }
941 return;
942 }
943 }
944
945 ret = ctdb_call_local(ctdb_db, call, &header, hdr, &data, true, c->hdr.srcnode);
946 if (ret != 0) {
947 DEBUG(DEBUG_ERR,(__location__ " ctdb_call_local failed\n"));
948 call->status = -1;
949 }
950
951 ret = ctdb_ltdb_unlock(ctdb_db, call->key);
952 if (ret != 0) {
953 DEBUG(DEBUG_ERR,(__location__ " ctdb_ltdb_unlock() failed with error %d\n", ret));
954 }
955
956 len = offsetof(struct ctdb_reply_call, data) + call->reply_data.dsize;
957 r = ctdb_transport_allocate(ctdb, ctdb, CTDB_REPLY_CALL, len,
958 struct ctdb_reply_call);
959 CTDB_NO_MEMORY_FATAL(ctdb, r);
960 r->hdr.destnode = hdr->srcnode;
961 r->hdr.reqid = hdr->reqid;
962 r->status = call->status;
963 r->datalen = call->reply_data.dsize;
964 if (call->reply_data.dsize) {
965 memcpy(&r->data[0], call->reply_data.dptr, call->reply_data.dsize);
966 }
967
968 ctdb_queue_packet(ctdb, &r->hdr);
969
970 talloc_free(r);
971 }
972
973 /*
974 called when a CTDB_REPLY_CALL packet comes in
975
976 This packet comes in response to a CTDB_REQ_CALL request packet. It
977 contains any reply data from the call
978 */
979 void ctdb_reply_call(struct ctdb_context *ctdb, struct ctdb_req_header *hdr)
980 {
981 struct ctdb_reply_call *c = (struct ctdb_reply_call *)hdr;
982 struct ctdb_call_state *state;
983
984 state = ctdb_reqid_find(ctdb, hdr->reqid, struct ctdb_call_state);
985 if (state == NULL) {
986 DEBUG(DEBUG_ERR, (__location__ " reqid %u not found\n", hdr->reqid));
987 return;
988 }
989
990 if (hdr->reqid != state->reqid) {
991 /* we found a record but it was the wrong one */
992 DEBUG(DEBUG_ERR, ("Dropped orphaned call reply with reqid:%u\n",hdr->reqid));
993 return;
994 }
995
996
997 /* read only delegation processing */
998 /* If we got a FETCH_WITH_HEADER we should check if this is a ro
999 * delegation since we may need to update the record header
1000 */
1001 if (state->c->callid == CTDB_FETCH_WITH_HEADER_FUNC) {
1002 struct ctdb_db_context *ctdb_db = state->ctdb_db;
1003 struct ctdb_ltdb_header *header = (struct ctdb_ltdb_header *)&c->data[0];
1004 struct ctdb_ltdb_header oldheader;
1005 TDB_DATA key, data, olddata;
1006 int ret;
1007
1008 if (!(header->flags & CTDB_REC_RO_HAVE_READONLY)) {
1009 goto finished_ro;
1010 return;
1011 }
1012
1013 key.dsize = state->c->keylen;
1014 key.dptr = state->c->data;
1015 ret = ctdb_ltdb_lock_requeue(ctdb_db, key, hdr,
1016 ctdb_call_input_pkt, ctdb, false);
1017 if (ret == -2) {
1018 return;
1019 }
1020 if (ret != 0) {
1021 DEBUG(DEBUG_ERR,(__location__ " Failed to get lock in ctdb_reply_call\n"));
1022 return;
1023 }
1024
1025 ret = ctdb_ltdb_fetch(ctdb_db, key, &oldheader, state, &olddata);
1026 if (ret != 0) {
1027 DEBUG(DEBUG_ERR, ("Failed to fetch old record in ctdb_reply_call\n"));
1028 ctdb_ltdb_unlock(ctdb_db, key);
1029 goto finished_ro;
1030 }
1031
1032 if (header->rsn <= oldheader.rsn) {
1033 ctdb_ltdb_unlock(ctdb_db, key);
1034 goto finished_ro;
1035 }
1036
1037 if (c->datalen < sizeof(struct ctdb_ltdb_header)) {
1038 DEBUG(DEBUG_ERR,(__location__ " Got FETCH_WITH_HEADER reply with too little data: %d bytes\n", c->datalen));
1039 ctdb_ltdb_unlock(ctdb_db, key);
1040 goto finished_ro;
1041 }
1042
1043 data.dsize = c->datalen - sizeof(struct ctdb_ltdb_header);
1044 data.dptr = &c->data[sizeof(struct ctdb_ltdb_header)];
1045 ret = ctdb_ltdb_store(ctdb_db, key, header, data);
1046 if (ret != 0) {
1047 DEBUG(DEBUG_ERR, ("Failed to store new record in ctdb_reply_call\n"));
1048 ctdb_ltdb_unlock(ctdb_db, key);
1049 goto finished_ro;
1050 }
1051
1052 ctdb_ltdb_unlock(ctdb_db, key);
1053 }
1054 finished_ro:
1055
1056 state->call->reply_data.dptr = c->data;
1057 state->call->reply_data.dsize = c->datalen;
1058 state->call->status = c->status;
1059
1060 talloc_steal(state, c);
1061
1062 state->state = CTDB_CALL_DONE;
1063 if (state->async.fn) {
1064 state->async.fn(state);
1065 }
1066 }
1067
1068
1069 /*
1070 called when a CTDB_REPLY_DMASTER packet comes in
1071
1072 This packet comes in from the lmaster response to a CTDB_REQ_CALL
1073 request packet. It means that the current dmaster wants to give us
1074 the dmaster role
1075 */
1076 void ctdb_reply_dmaster(struct ctdb_context *ctdb, struct ctdb_req_header *hdr)
1077 {
1078 struct ctdb_reply_dmaster *c = (struct ctdb_reply_dmaster *)hdr;
1079 struct ctdb_db_context *ctdb_db;
1080 TDB_DATA key, data;
1081 uint32_t record_flags = 0;
1082 size_t len;
1083 int ret;
1084
1085 ctdb_db = find_ctdb_db(ctdb, c->db_id);
1086 if (ctdb_db == NULL) {
1087 DEBUG(DEBUG_ERR,("Unknown db_id 0x%x in ctdb_reply_dmaster\n", c->db_id));
1088 return;
1089 }
1090
1091 key.dptr = c->data;
1092 key.dsize = c->keylen;
1093 data.dptr = &c->data[key.dsize];
1094 data.dsize = c->datalen;
1095 len = offsetof(struct ctdb_reply_dmaster, data) + key.dsize + data.dsize
1096 + sizeof(uint32_t);
1097 if (len <= c->hdr.length) {
1098 record_flags = *(uint32_t *)&c->data[c->keylen + c->datalen];
1099 }
1100
1101 ret = ctdb_ltdb_lock_requeue(ctdb_db, key, hdr,
1102 ctdb_call_input_pkt, ctdb, false);
1103 if (ret == -2) {
1104 return;
1105 }
1106 if (ret != 0) {
1107 DEBUG(DEBUG_ERR,(__location__ " Failed to get lock in ctdb_reply_dmaster\n"));
1108 return;
1109 }
1110
1111 ctdb_become_dmaster(ctdb_db, hdr, key, data, c->rsn, record_flags);
1112 }
1113
1114
1115 /*
1116 called when a CTDB_REPLY_ERROR packet comes in
1117 */
1118 void ctdb_reply_error(struct ctdb_context *ctdb, struct ctdb_req_header *hdr)
1119 {
1120 struct ctdb_reply_error *c = (struct ctdb_reply_error *)hdr;
1121 struct ctdb_call_state *state;
1122
1123 state = ctdb_reqid_find(ctdb, hdr->reqid, struct ctdb_call_state);
1124 if (state == NULL) {
1125 DEBUG(DEBUG_ERR,("pnn %u Invalid reqid %u in ctdb_reply_error\n",
1126 ctdb->pnn, hdr->reqid));
1127 return;
1128 }
1129
1130 if (hdr->reqid != state->reqid) {
1131 /* we found a record but it was the wrong one */
1132 DEBUG(DEBUG_ERR, ("Dropped orphaned error reply with reqid:%u\n",hdr->reqid));
1133 return;
1134 }
1135
1136 talloc_steal(state, c);
1137
1138 state->state = CTDB_CALL_ERROR;
1139 state->errmsg = (char *)c->msg;
1140 if (state->async.fn) {
1141 state->async.fn(state);
1142 }
1143 }
1144
1145
1146 /*
1147 destroy a ctdb_call
1148 */
1149 static int ctdb_call_destructor(struct ctdb_call_state *state)
1150 {
1151 DLIST_REMOVE(state->ctdb_db->ctdb->pending_calls, state);
1152 ctdb_reqid_remove(state->ctdb_db->ctdb, state->reqid);
1153 return 0;
1154 }
1155
1156
1157 /*
1158 called when a ctdb_call needs to be resent after a reconfigure event
1159 */
1160 static void ctdb_call_resend(struct ctdb_call_state *state)
1161 {
1162 struct ctdb_context *ctdb = state->ctdb_db->ctdb;
1163
1164 state->generation = ctdb->vnn_map->generation;
1165
1166 /* use a new reqid, in case the old reply does eventually come in */
1167 ctdb_reqid_remove(ctdb, state->reqid);
1168 state->reqid = ctdb_reqid_new(ctdb, state);
1169 state->c->hdr.reqid = state->reqid;
1170
1171 /* update the generation count for this request, so its valid with the new vnn_map */
1172 state->c->hdr.generation = state->generation;
1173
1174 /* send the packet to ourselves, it will be redirected appropriately */
1175 state->c->hdr.destnode = ctdb->pnn;
1176
1177 ctdb_queue_packet(ctdb, &state->c->hdr);
1178 DEBUG(DEBUG_NOTICE,("resent ctdb_call\n"));
1179 }
1180
1181 /*
1182 resend all pending calls on recovery
1183 */
1184 void ctdb_call_resend_all(struct ctdb_context *ctdb)
1185 {
1186 struct ctdb_call_state *state, *next;
1187 for (state=ctdb->pending_calls;state;state=next) {
1188 next = state->next;
1189 ctdb_call_resend(state);
1190 }
1191 }
1192
1193 /*
1194 this allows the caller to setup a async.fn
1195 */
1196 static void call_local_trigger(struct event_context *ev, struct timed_event *te,
1197 struct timeval t, void *private_data)
1198 {
1199 struct ctdb_call_state *state = talloc_get_type(private_data, struct ctdb_call_state);
1200 if (state->async.fn) {
1201 state->async.fn(state);
1202 }
1203 }
1204
1205
1206 /*
1207 construct an event driven local ctdb_call
1208
1209 this is used so that locally processed ctdb_call requests are processed
1210 in an event driven manner
1211 */
1212 struct ctdb_call_state *ctdb_call_local_send(struct ctdb_db_context *ctdb_db,
1213 struct ctdb_call *call,
1214 struct ctdb_ltdb_header *header,
1215 TDB_DATA *data)
1216 {
1217 struct ctdb_call_state *state;
1218 struct ctdb_context *ctdb = ctdb_db->ctdb;
1219 int ret;
1220
1221 state = talloc_zero(ctdb_db, struct ctdb_call_state);
1222 CTDB_NO_MEMORY_NULL(ctdb, state);
1223
1224 talloc_steal(state, data->dptr);
1225
1226 state->state = CTDB_CALL_DONE;
1227 state->call = talloc(state, struct ctdb_call);
1228 CTDB_NO_MEMORY_NULL(ctdb, state->call);
1229 *(state->call) = *call;
1230 state->ctdb_db = ctdb_db;
1231
1232 ret = ctdb_call_local(ctdb_db, state->call, header, state, data, true, ctdb->pnn);
1233 if (ret != 0) {
1234 DEBUG(DEBUG_DEBUG,("ctdb_call_local() failed, ignoring return code %d\n", ret));
1235 }
1236
1237 event_add_timed(ctdb->ev, state, timeval_zero(), call_local_trigger, state);
1238
1239 return state;
1240 }
1241
1242
1243 /*
1244 make a remote ctdb call - async send. Called in daemon context.
1245
1246 This constructs a ctdb_call request and queues it for processing.
1247 This call never blocks.
1248 */
1249 struct ctdb_call_state *ctdb_daemon_call_send_remote(struct ctdb_db_context *ctdb_db,
1250 struct ctdb_call *call,
1251 struct ctdb_ltdb_header *header)
1252 {
1253 uint32_t len;
1254 struct ctdb_call_state *state;
1255 struct ctdb_context *ctdb = ctdb_db->ctdb;
1256
1257 if (ctdb->methods == NULL) {
1258 DEBUG(DEBUG_INFO,(__location__ " Failed send packet. Transport is down\n"));
1259 return NULL;
1260 }
1261
1262 state = talloc_zero(ctdb_db, struct ctdb_call_state);
1263 CTDB_NO_MEMORY_NULL(ctdb, state);
1264 state->call = talloc(state, struct ctdb_call);
1265 CTDB_NO_MEMORY_NULL(ctdb, state->call);
1266
1267 state->reqid = ctdb_reqid_new(ctdb, state);
1268 state->ctdb_db = ctdb_db;
1269 talloc_set_destructor(state, ctdb_call_destructor);
1270
1271 len = offsetof(struct ctdb_req_call, data) + call->key.dsize + call->call_data.dsize;
1272 state->c = ctdb_transport_allocate(ctdb, state, CTDB_REQ_CALL, len,
1273 struct ctdb_req_call);
1274 CTDB_NO_MEMORY_NULL(ctdb, state->c);
1275 state->c->hdr.destnode = header->dmaster;
1276
1277 /* this limits us to 16k outstanding messages - not unreasonable */
1278 state->c->hdr.reqid = state->reqid;
1279 state->c->flags = call->flags;
1280 state->c->db_id = ctdb_db->db_id;
1281 state->c->callid = call->call_id;
1282 state->c->hopcount = 0;
1283 state->c->keylen = call->key.dsize;
1284 state->c->calldatalen = call->call_data.dsize;
1285 memcpy(&state->c->data[0], call->key.dptr, call->key.dsize);
1286 memcpy(&state->c->data[call->key.dsize],
1287 call->call_data.dptr, call->call_data.dsize);
1288 *(state->call) = *call;
1289 state->call->call_data.dptr = &state->c->data[call->key.dsize];
1290 state->call->key.dptr = &state->c->data[0];
1291
1292 state->state = CTDB_CALL_WAIT;
1293 state->generation = ctdb->vnn_map->generation;
1294
1295 DLIST_ADD(ctdb->pending_calls, state);
1296
1297 ctdb_queue_packet(ctdb, &state->c->hdr);
1298
1299 return state;
1300 }
1301
1302 /*
1303 make a remote ctdb call - async recv - called in daemon context
1304
1305 This is called when the program wants to wait for a ctdb_call to complete and get the
1306 results. This call will block unless the call has already completed.
1307 */
1308 int ctdb_daemon_call_recv(struct ctdb_call_state *state, struct ctdb_call *call)
1309 {
1310 while (state->state < CTDB_CALL_DONE) {
1311 event_loop_once(state->ctdb_db->ctdb->ev);
1312 }
1313 if (state->state != CTDB_CALL_DONE) {
1314 ctdb_set_error(state->ctdb_db->ctdb, "%s", state->errmsg);
1315 talloc_free(state);
1316 return -1;
1317 }
1318
1319 if (state->call->reply_data.dsize) {
1320 call->reply_data.dptr = talloc_memdup(call,
1321 state->call->reply_data.dptr,
1322 state->call->reply_data.dsize);
1323 call->reply_data.dsize = state->call->reply_data.dsize;
1324 } else {
1325 call->reply_data.dptr = NULL;
1326 call->reply_data.dsize = 0;
1327 }
1328 call->status = state->call->status;
1329 talloc_free(state);
1330 return 0;
1331 }
1332
1333
1334 /*
1335 send a keepalive packet to the other node
1336 */
1337 void ctdb_send_keepalive(struct ctdb_context *ctdb, uint32_t destnode)
1338 {
1339 struct ctdb_req_keepalive *r;
1340
1341 if (ctdb->methods == NULL) {
1342 DEBUG(DEBUG_INFO,(__location__ " Failed to send keepalive. Transport is DOWN\n"));
1343 return;
1344 }
1345
1346 r = ctdb_transport_allocate(ctdb, ctdb, CTDB_REQ_KEEPALIVE,
1347 sizeof(struct ctdb_req_keepalive),
1348 struct ctdb_req_keepalive);
1349 CTDB_NO_MEMORY_FATAL(ctdb, r);
1350 r->hdr.destnode = destnode;
1351 r->hdr.reqid = 0;
1352
1353 CTDB_INCREMENT_STAT(ctdb, keepalive_packets_sent);
1354
1355 ctdb_queue_packet(ctdb, &r->hdr);
1356
1357 talloc_free(r);
1358 }
1359
1360
1361
1362 struct revokechild_deferred_call {
1363 struct ctdb_context *ctdb;
1364 struct ctdb_req_header *hdr;
1365 deferred_requeue_fn fn;
1366 void *ctx;
1367 };
1368
1369 struct revokechild_handle {
1370 struct revokechild_handle *next, *prev;
1371 struct ctdb_context *ctdb;
1372 struct ctdb_db_context *ctdb_db;
1373 struct fd_event *fde;
1374 int status;
1375 int fd[2];
1376 pid_t child;
1377 TDB_DATA key;
1378 };
1379
1380 struct revokechild_requeue_handle {
1381 struct ctdb_context *ctdb;
1382 struct ctdb_req_header *hdr;
1383 deferred_requeue_fn fn;
1384 void *ctx;
1385 };
1386
1387 static void deferred_call_requeue(struct event_context *ev, struct timed_event *te,
1388 struct timeval t, void *private_data)
1389 {
1390 struct revokechild_requeue_handle *requeue_handle = talloc_get_type(private_data, struct revokechild_requeue_handle);
1391
1392 requeue_handle->fn(requeue_handle->ctx, requeue_handle->hdr);
1393 talloc_free(requeue_handle);
1394 }
1395
1396 static int deferred_call_destructor(struct revokechild_deferred_call *deferred_call)
1397 {
1398 struct ctdb_context *ctdb = deferred_call->ctdb;
1399 struct revokechild_requeue_handle *requeue_handle = talloc(ctdb, struct revokechild_requeue_handle);
1400 struct ctdb_req_call *c = (struct ctdb_req_call *)deferred_call->hdr;
1401
1402 requeue_handle->ctdb = ctdb;
1403 requeue_handle->hdr = deferred_call->hdr;
1404 requeue_handle->fn = deferred_call->fn;
1405 requeue_handle->ctx = deferred_call->ctx;
1406 talloc_steal(requeue_handle, requeue_handle->hdr);
1407
1408 /* when revoking, any READONLY requests have 1 second grace to let read/write finish first */
1409 event_add_timed(ctdb->ev, requeue_handle, timeval_current_ofs(c->flags & CTDB_WANT_READONLY ? 1 : 0, 0), deferred_call_requeue, requeue_handle);
1410
1411 return 0;
1412 }
1413
1414
1415 static int revokechild_destructor(struct revokechild_handle *rc)
1416 {
1417 if (rc->fde != NULL) {
1418 talloc_free(rc->fde);
1419 }
1420
1421 if (rc->fd[0] != -1) {
1422 close(rc->fd[0]);
1423 }
1424 if (rc->fd[1] != -1) {
1425 close(rc->fd[1]);
1426 }
1427 ctdb_kill(rc->ctdb, rc->child, SIGKILL);
1428
1429 DLIST_REMOVE(rc->ctdb_db->revokechild_active, rc);
1430 return 0;
1431 }
1432
1433 static void revokechild_handler(struct event_context *ev, struct fd_event *fde,
1434 uint16_t flags, void *private_data)
1435 {
1436 struct revokechild_handle *rc = talloc_get_type(private_data,
1437 struct revokechild_handle);
1438 int ret;
1439 char c;
1440
1441 ret = read(rc->fd[0], &c, 1);
1442 if (ret != 1) {
1443 DEBUG(DEBUG_ERR,("Failed to read status from revokechild. errno:%d\n", errno));
1444 rc->status = -1;
1445 talloc_free(rc);
1446 return;
1447 }
1448 if (c != 0) {
1449 DEBUG(DEBUG_ERR,("revokechild returned failure. status:%d\n", c));
1450 rc->status = -1;
1451 talloc_free(rc);
1452 return;
1453 }
1454
1455 talloc_free(rc);
1456 }
1457
1458 struct ctdb_revoke_state {
1459 struct ctdb_db_context *ctdb_db;
1460 TDB_DATA key;
1461 struct ctdb_ltdb_header *header;
1462 TDB_DATA data;
1463 int count;
1464 int status;
1465 int finished;
1466 };
1467
1468 static void update_record_cb(struct ctdb_client_control_state *state)
1469 {
1470 struct ctdb_revoke_state *revoke_state;
1471 int ret;
1472 int32_t res;
1473
1474 if (state == NULL) {
1475 return;
1476 }
1477 revoke_state = state->async.private_data;
1478
1479 state->async.fn = NULL;
1480 ret = ctdb_control_recv(state->ctdb, state, state, NULL, &res, NULL);
1481 if ((ret != 0) || (res != 0)) {
1482 DEBUG(DEBUG_ERR,("Recv for revoke update record failed ret:%d res:%d\n", ret, res));
1483 revoke_state->status = -1;
1484 }
1485
1486 revoke_state->count--;
1487 if (revoke_state->count <= 0) {
1488 revoke_state->finished = 1;
1489 }
1490 }
1491
1492 static void revoke_send_cb(struct ctdb_context *ctdb, uint32_t pnn, void *private_data)
1493 {
1494 struct ctdb_revoke_state *revoke_state = private_data;
1495 struct ctdb_client_control_state *state;
1496
1497 state = ctdb_ctrl_updaterecord_send(ctdb, revoke_state, timeval_current_ofs(5,0), pnn, revoke_state->ctdb_db, revoke_state->key, revoke_state->header, revoke_state->data);
1498 if (state == NULL) {
1499 DEBUG(DEBUG_ERR,("Failure to send update record to revoke readonly delegation\n"));
1500 revoke_state->status = -1;
1501 return;
1502 }
1503 state->async.fn = update_record_cb;
1504 state->async.private_data = revoke_state;
1505
1506 revoke_state->count++;
1507
1508 }
1509
1510 static void ctdb_revoke_timeout_handler(struct event_context *ev, struct timed_event *te,
1511 struct timeval yt, void *private_data)
1512 {
1513 struct ctdb_revoke_state *state = private_data;
1514
1515 DEBUG(DEBUG_ERR,("Timed out waiting for revoke to finish\n"));
1516 state->finished = 1;
1517 state->status = -1;
1518 }
1519
1520 static int ctdb_revoke_all_delegations(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, TDB_DATA tdata, TDB_DATA key, struct ctdb_ltdb_header *header, TDB_DATA data)
1521 {
1522 struct ctdb_revoke_state *state = talloc_zero(ctdb, struct ctdb_revoke_state);
1523 int status;
1524
1525 state->ctdb_db = ctdb_db;
1526 state->key = key;
1527 state->header = header;
1528 state->data = data;
1529
1530 ctdb_trackingdb_traverse(ctdb, tdata, revoke_send_cb, state);
1531
1532 event_add_timed(ctdb->ev, state, timeval_current_ofs(5, 0), ctdb_revoke_timeout_handler, state);
1533
1534 while (state->finished == 0) {
1535 event_loop_once(ctdb->ev);
1536 }
1537
1538 status = state->status;
1539
1540 if (status == 0) {
1541 struct ctdb_ltdb_header new_header;
1542 TDB_DATA new_data;
1543
1544 if (ctdb_ltdb_lock(ctdb_db, key) != 0) {
1545 DEBUG(DEBUG_ERR,("Failed to chainlock the database in revokechild\n"));
1546 talloc_free(state);
1547 return -1;
1548 }
1549 if (ctdb_ltdb_fetch(ctdb_db, key, &new_header, state, &new_data) != 0) {
1550 ctdb_ltdb_unlock(ctdb_db, key);
1551 DEBUG(DEBUG_ERR,("Failed for fetch tdb record in revokechild\n"));
1552 talloc_free(state);
1553 return -1;
1554 }
1555 header->rsn++;
1556 if (new_header.rsn > header->rsn) {
1557 ctdb_ltdb_unlock(ctdb_db, key);
1558 DEBUG(DEBUG_ERR,("RSN too high in tdb record in revokechild\n"));
1559 talloc_free(state);
1560 return -1;
1561 }
1562 if ( (new_header.flags & (CTDB_REC_RO_REVOKING_READONLY|CTDB_REC_RO_HAVE_DELEGATIONS)) != (CTDB_REC_RO_REVOKING_READONLY|CTDB_REC_RO_HAVE_DELEGATIONS) ) {
1563 ctdb_ltdb_unlock(ctdb_db, key);
1564 DEBUG(DEBUG_ERR,("Flags are wrong in tdb record in revokechild\n"));
1565 talloc_free(state);
1566 return -1;
1567 }
1568 new_header.rsn++;
1569 new_header.flags |= CTDB_REC_RO_REVOKE_COMPLETE;
1570 if (ctdb_ltdb_store(ctdb_db, key, &new_header, new_data) != 0) {
1571 ctdb_ltdb_unlock(ctdb_db, key);
1572 DEBUG(DEBUG_ERR,("Failed to write new record in revokechild\n"));
1573 talloc_free(state);
1574 return -1;
1575 }
1576 ctdb_ltdb_unlock(ctdb_db, key);
1577 }
1578
1579 talloc_free(state);
1580 return status;
1581 }
1582
1583
1584 int ctdb_start_revoke_ro_record(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, TDB_DATA key, struct ctdb_ltdb_header *header, TDB_DATA data)
1585 {
1586 TDB_DATA tdata;
1587 struct revokechild_handle *rc;
1588 pid_t parent = getpid();
1589 int ret;
1590
1591 header->flags &= ~(CTDB_REC_RO_REVOKING_READONLY|CTDB_REC_RO_HAVE_DELEGATIONS|CTDB_REC_RO_HAVE_READONLY);
1592 header->flags |= CTDB_REC_FLAG_MIGRATED_WITH_DATA;
1593 header->rsn -= 1;
1594
1595 if ((rc = talloc_zero(ctdb_db, struct revokechild_handle)) == NULL) {
1596 DEBUG(DEBUG_ERR,("Failed to allocate revokechild_handle\n"));
1597 return -1;
1598 }
1599
1600 tdata = tdb_fetch(ctdb_db->rottdb, key);
1601 if (tdata.dsize > 0) {
1602 uint8_t *tmp;
1603
1604 tmp = tdata.dptr;
1605 tdata.dptr = talloc_memdup(rc, tdata.dptr, tdata.dsize);
1606 free(tmp);
1607 }
1608
1609 rc->status = 0;
1610 rc->ctdb = ctdb;
1611 rc->ctdb_db = ctdb_db;
1612 rc->fd[0] = -1;
1613 rc->fd[1] = -1;
1614
1615 talloc_set_destructor(rc, revokechild_destructor);
1616
1617 rc->key.dsize = key.dsize;
1618 rc->key.dptr = talloc_memdup(rc, key.dptr, key.dsize);
1619 if (rc->key.dptr == NULL) {
1620 DEBUG(DEBUG_ERR,("Failed to allocate key for revokechild_handle\n"));
1621 talloc_free(rc);
1622 return -1;
1623 }
1624
1625 ret = pipe(rc->fd);
1626 if (ret != 0) {
1627 DEBUG(DEBUG_ERR,("Failed to allocate key for revokechild_handle\n"));
1628 talloc_free(rc);
1629 return -1;
1630 }
1631
1632
1633 rc->child = ctdb_fork(ctdb);
1634 if (rc->child == (pid_t)-1) {
1635 DEBUG(DEBUG_ERR,("Failed to fork child for revokechild\n"));
1636 talloc_free(rc);
1637 return -1;
1638 }
1639
1640 if (rc->child == 0) {
1641 char c = 0;
1642 close(rc->fd[0]);
1643 debug_extra = talloc_asprintf(NULL, "revokechild-%s:", ctdb_db->db_name);
1644
1645 if (switch_from_server_to_client(ctdb, "revokechild-%s", ctdb_db->db_name) != 0) {
1646 DEBUG(DEBUG_ERR,("Failed to switch from server to client for revokechild process\n"));
1647 c = 1;
1648 goto child_finished;
1649 }
1650
1651 c = ctdb_revoke_all_delegations(ctdb, ctdb_db, tdata, key, header, data);
1652
1653 child_finished:
1654 write(rc->fd[1], &c, 1);
1655 /* make sure we die when our parent dies */
1656 while (ctdb_kill(ctdb, parent, 0) == 0 || errno != ESRCH) {
1657 sleep(5);
1658 }
1659 _exit(0);
1660 }
1661
1662 close(rc->fd[1]);
1663 rc->fd[1] = -1;
1664 set_close_on_exec(rc->fd[0]);
1665
1666 /* This is an active revokechild child process */
1667 DLIST_ADD_END(ctdb_db->revokechild_active, rc, NULL);
1668
1669 rc->fde = event_add_fd(ctdb->ev, rc, rc->fd[0],
1670 EVENT_FD_READ, revokechild_handler,
1671 (void *)rc);
1672 if (rc->fde == NULL) {
1673 DEBUG(DEBUG_ERR,("Failed to set up fd event for revokechild process\n"));
1674 talloc_free(rc);
1675 }
1676 tevent_fd_set_auto_close(rc->fde);
1677
1678 return 0;
1679 }
1680
1681 int ctdb_add_revoke_deferred_call(struct ctdb_context *ctdb, struct ctdb_db_context *ctdb_db, TDB_DATA key, struct ctdb_req_header *hdr, deferred_requeue_fn fn, void *call_context)
1682 {
1683 struct revokechild_handle *rc;
1684 struct revokechild_deferred_call *deferred_call;
1685
1686 for (rc = ctdb_db->revokechild_active; rc; rc = rc->next) {
1687 if (rc->key.dsize == 0) {
1688 continue;
1689 }
1690 if (rc->key.dsize != key.dsize) {
1691 continue;
1692 }
1693 if (!memcmp(rc->key.dptr, key.dptr, key.dsize)) {
1694 break;
1695 }
1696 }
1697
1698 if (rc == NULL) {
1699 DEBUG(DEBUG_ERR,("Failed to add deferred call to revoke list. revoke structure not found\n"));
1700 return -1;
1701 }
1702
1703 deferred_call = talloc(rc, struct revokechild_deferred_call);
1704 if (deferred_call == NULL) {
1705 DEBUG(DEBUG_ERR,("Failed to allocate deferred call structure for revoking record\n"));
1706 return -1;
1707 }
1708
1709 deferred_call->ctdb = ctdb;
1710 deferred_call->hdr = hdr;
1711 deferred_call->fn = fn;
1712 deferred_call->ctx = call_context;
1713
1714 talloc_set_destructor(deferred_call, deferred_call_destructor);
1715 talloc_steal(deferred_call, hdr);
1716
1717 return 0;
1718 }
Samba GIT mirror