5 * DEBUG: section 22 Refresh Calculation
6 * AUTHOR: Harvest Derived
8 * SQUID Web Proxy Cache http://www.squid-cache.org/
9 * ----------------------------------------------------------
11 * Squid is the result of efforts by numerous individuals from
12 * the Internet community; see the CONTRIBUTORS file for full
13 * details. Many organizations have provided support for Squid's
14 * development; see the SPONSORS file for full details. Squid is
15 * Copyrighted (C) 2001 by the Regents of the University of
16 * California; see the COPYRIGHT file for full details. Squid
17 * incorporates software developed and/or copyrighted by other
18 * sources; see the CREDITS file for full details.
20 * This program is free software; you can redistribute it and/or modify
21 * it under the terms of the GNU General Public License as published by
22 * the Free Software Foundation; either version 2 of the License, or
23 * (at your option) any later version.
25 * This program is distributed in the hope that it will be useful,
26 * but WITHOUT ANY WARRANTY; without even the implied warranty of
27 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
28 * GNU General Public License for more details.
30 * You should have received a copy of the GNU General Public License
31 * along with this program; if not, write to the Free Software
32 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
36 #ifndef USE_POSIX_REGEX
37 #define USE_POSIX_REGEX /* put before includes; always use POSIX */
41 #include "CacheManager.h"
43 #include "MemObject.h"
44 #include "HttpRequest.h"
45 #include "HttpReply.h"
46 #include "SquidTime.h"
69 * This enumerated list assigns specific values, ala HTTP/FTP status
70 * codes. All Fresh codes are in the range 100-199 and all stale
71 * codes are 200-299. We might want to use these codes in logging,
72 * so best to keep them consistent over time.
75 FRESH_REQUEST_MAX_STALE_ALL
= 100,
76 FRESH_REQUEST_MAX_STALE_VALUE
,
80 FRESH_OVERRIDE_EXPIRES
,
81 FRESH_OVERRIDE_LASTMOD
,
82 STALE_MUST_REVALIDATE
= 200,
83 STALE_RELOAD_INTO_IMS
,
85 STALE_EXCEEDS_REQUEST_MAX_AGE_VALUE
,
92 static struct RefreshCounts
{
95 int status
[STALE_DEFAULT
+ 1];
98 refreshCounts
[rcCount
];
106 #define REFRESH_DEFAULT_MIN (time_t)0
107 #define REFRESH_DEFAULT_PCT 0.20
108 #define REFRESH_DEFAULT_MAX (time_t)259200
110 static const refresh_t
*refreshUncompiledPattern(const char *);
111 static OBJH refreshStats
;
112 static int refreshStaleness(const StoreEntry
* entry
, time_t check_time
, const time_t age
, const refresh_t
* R
, stale_flags
* sf
);
114 static refresh_t DefaultRefresh
;
117 refreshLimits(const char *url
)
121 for (R
= Config
.Refresh
; R
; R
= R
->next
) {
122 if (!regexec(&(R
->compiled_pattern
), url
, 0, 0, 0))
129 static const refresh_t
*
130 refreshUncompiledPattern(const char *pat
)
134 for (R
= Config
.Refresh
; R
; R
= R
->next
) {
135 if (0 == strcmp(R
->pattern
, pat
))
143 * Calculate how stale the response is (or will be at the check_time).
144 * Staleness calculation is based on the following: (1) response
145 * expiration time, (2) age greater than configured maximum, (3)
146 * last-modified factor, and (4) age less than configured minimum.
148 * \retval -1 If the response is fresh.
149 * \retval >0 Otherwise return it's staleness.
150 * \retval 0 NOTE return value of 0 means the response is stale.
152 * The 'stale_flags' structure is used to tell the calling function
153 * _why_ this response is fresh or stale. Its used, for example,
154 * when the admin wants to override expiration and last-modified
158 refreshStaleness(const StoreEntry
* entry
, time_t check_time
, const time_t age
, const refresh_t
* R
, stale_flags
* sf
)
161 * Check for an explicit expiration time (Expires: header).
163 if (entry
->expires
> -1) {
166 if (entry
->expires
> check_time
) {
167 debugs(22, 3, "FRESH: expires " << entry
->expires
<<
168 " >= check_time " << check_time
<< " ");
172 debugs(22, 3, "STALE: expires " << entry
->expires
<<
173 " < check_time " << check_time
<< " ");
175 return (check_time
- entry
->expires
);
180 * Use local heuristics to determine staleness. Start with the
181 * max age from the refresh_pattern rule.
184 debugs(22, 3, "STALE: age " << age
<< " > max " << R
->max
<< " ");
186 return (age
- R
->max
);
190 * Try the last-modified factor algorithm: refresh_pattern n% percentage of Last-Modified: age.
192 if (entry
->lastmod
> -1 && entry
->timestamp
> entry
->lastmod
) {
194 * stale_age is the Age of the response when it became/becomes
195 * stale according to the last-modified factor algorithm.
197 time_t stale_age
= static_cast<time_t>((entry
->timestamp
- entry
->lastmod
) * R
->pct
);
200 if (age
>= stale_age
) {
201 debugs(22, 3, "STALE: age " << age
<< " > stale_age " << stale_age
);
202 return (age
- stale_age
);
204 debugs(22, 3, "FRESH: age " << age
<< " <= stale_age " << stale_age
);
210 * Finally, if all else fails; staleness is determined by the refresh_pattern
211 * configured minimum age.
214 debugs(22, 3, "FRESH: age " << age
<< " < min " << R
->min
);
219 debugs(22, 3, "STALE: age " << age
<< " >= min " << R
->min
);
220 return (age
- R
->min
);
224 * \retval 1 if the entry must be revalidated within delta seconds
225 * \retval 0 otherwise
227 * note: request maybe null (e.g. for cache digests build)
230 refreshCheck(const StoreEntry
* entry
, HttpRequest
* request
, time_t delta
)
233 const char *uri
= NULL
;
235 time_t check_time
= squid_curtime
+ delta
;
240 uri
= entry
->mem_obj
->url
;
242 uri
= urlCanonical(request
);
244 debugs(22, 3, "refreshCheck: '" << (uri
? uri
: "<none>") << "'");
246 if (check_time
> entry
->timestamp
)
247 age
= check_time
- entry
->timestamp
;
249 // FIXME: what to do when age < 0 or counter overflow?
252 R
= uri
? refreshLimits(uri
) : refreshUncompiledPattern(".");
257 memset(&sf
, '\0', sizeof(sf
));
259 staleness
= refreshStaleness(entry
, check_time
, age
, R
, &sf
);
261 debugs(22, 3, "Staleness = " << staleness
);
263 debugs(22, 3, "refreshCheck: Matched '" << R
->pattern
<< " " <<
264 (int) R
->min
<< " " << (int) (100.0 * R
->pct
) << "%% " <<
265 (int) R
->max
<< "'");
268 debugs(22, 3, "refreshCheck: age = " << age
);
270 debugs(22, 3, "\tcheck_time:\t" << mkrfc1123(check_time
));
272 debugs(22, 3, "\tentry->timestamp:\t" << mkrfc1123(entry
->timestamp
));
274 if (EBIT_TEST(entry
->flags
, ENTRY_REVALIDATE
) && staleness
> -1
275 #if USE_HTTP_VIOLATIONS
276 && !R
->flags
.ignore_must_revalidate
279 debugs(22, 3, "refreshCheck: YES: Must revalidate stale response");
280 return STALE_MUST_REVALIDATE
;
283 /* request-specific checks */
284 if (request
&& !request
->flags
.ignore_cc
) {
285 HttpHdrCc
*cc
= request
->cache_control
;
287 if (request
->flags
.ims
&& (R
->flags
.refresh_ims
|| Config
.onoff
.refresh_all_ims
)) {
288 /* The clients no-cache header is changed into a IMS query */
289 debugs(22, 3, "refreshCheck: YES: refresh-ims");
290 return STALE_FORCED_RELOAD
;
293 #if USE_HTTP_VIOLATIONS
295 if (!request
->flags
.nocache_hack
) {
297 } else if (R
->flags
.ignore_reload
) {
298 /* The clients no-cache header is ignored */
299 debugs(22, 3, "refreshCheck: MAYBE: ignore-reload");
300 } else if (R
->flags
.reload_into_ims
|| Config
.onoff
.reload_into_ims
) {
301 /* The clients no-cache header is changed into a IMS query */
302 debugs(22, 3, "refreshCheck: YES: reload-into-ims");
303 return STALE_RELOAD_INTO_IMS
;
305 /* The clients no-cache header is not overridden on this request */
306 debugs(22, 3, "refreshCheck: YES: client reload");
307 request
->flags
.nocache
= 1;
308 return STALE_FORCED_RELOAD
;
313 if (cc
->max_age
> -1) {
314 #if USE_HTTP_VIOLATIONS
315 if (R
->flags
.ignore_reload
&& cc
->max_age
== 0) {} else
320 if (cc
->max_age
== 0) {
321 debugs(22, 3, "refreshCheck: YES: client-max-age = 0");
322 return STALE_EXCEEDS_REQUEST_MAX_AGE_VALUE
;
326 if (age
> cc
->max_age
) {
327 debugs(22, 3, "refreshCheck: YES: age > client-max-age");
328 return STALE_EXCEEDS_REQUEST_MAX_AGE_VALUE
;
333 if (EBIT_TEST(cc
->mask
, CC_MAX_STALE
) && staleness
> -1) {
334 if (cc
->max_stale
< 0) {
335 /* max-stale directive without a value */
336 debugs(22, 3, "refreshCheck: NO: max-stale wildcard");
337 return FRESH_REQUEST_MAX_STALE_ALL
;
338 } else if (staleness
< cc
->max_stale
) {
339 debugs(22, 3, "refreshCheck: NO: staleness < max-stale");
340 return FRESH_REQUEST_MAX_STALE_VALUE
;
346 if (-1 == staleness
) {
347 debugs(22, 3, "refreshCheck: object isn't stale..");
349 debugs(22, 3, "refreshCheck: returning FRESH_EXPIRES");
350 return FRESH_EXPIRES
;
356 debugs(22, 3, "refreshCheck: returning FRESH_LMFACTOR_RULE");
357 return FRESH_LMFACTOR_RULE
;
362 debugs(22, 3, "refreshCheck: returning FRESH_MIN_RULE");
363 return FRESH_MIN_RULE
;
367 * At this point the response is stale, unless one of
368 * the override options kicks in.
371 #if USE_HTTP_VIOLATIONS
373 if (R
->flags
.override_expire
&& age
< R
->min
) {
374 debugs(22, 3, "refreshCheck: NO: age < min && override-expire");
375 return FRESH_OVERRIDE_EXPIRES
;
379 return STALE_EXPIRES
;
383 return STALE_MAX_RULE
;
386 #if USE_HTTP_VIOLATIONS
388 if (R
->flags
.override_lastmod
&& age
< R
->min
) {
389 debugs(22, 3, "refreshCheck: NO: age < min && override-lastmod");
390 return FRESH_OVERRIDE_LASTMOD
;
394 return STALE_LMFACTOR_RULE
;
397 debugs(22, 3, "refreshCheck: returning STALE_DEFAULT");
398 return STALE_DEFAULT
;
402 refreshIsCachable(const StoreEntry
* entry
)
405 * Don't look at the request to avoid no-cache and other nuisances.
406 * the object should have a mem_obj so the URL will be found there.
407 * minimum_expiry_time seconds delta (defaults to 60 seconds), to
408 * avoid objects which expire almost immediately, and which can't
411 int reason
= refreshCheck(entry
, NULL
, Config
.minimum_expiry_time
);
412 refreshCounts
[rcStore
].total
++;
413 refreshCounts
[rcStore
].status
[reason
]++;
415 if (reason
< STALE_MUST_REVALIDATE
)
416 /* Does not need refresh. This is certainly cachable */
419 if (entry
->lastmod
< 0)
420 /* Last modified is needed to do a refresh */
423 if (entry
->mem_obj
== NULL
)
427 if (entry
->getReply() == NULL
)
431 if (entry
->getReply()->content_length
== 0)
432 /* No use refreshing (caching?) 0 byte objects */
435 /* This seems to be refreshable. Cache it */
439 /* refreshCheck... functions below are protocol-specific wrappers around
440 * refreshCheck() function above */
443 refreshCheckHTTP(const StoreEntry
* entry
, HttpRequest
* request
)
445 int reason
= refreshCheck(entry
, request
, 0);
446 refreshCounts
[rcHTTP
].total
++;
447 refreshCounts
[rcHTTP
].status
[reason
]++;
448 return (reason
< 200) ? 0 : 1;
452 refreshCheckICP(const StoreEntry
* entry
, HttpRequest
* request
)
454 int reason
= refreshCheck(entry
, request
, 30);
455 refreshCounts
[rcICP
].total
++;
456 refreshCounts
[rcICP
].status
[reason
]++;
457 return (reason
< 200) ? 0 : 1;
462 refreshCheckHTCP(const StoreEntry
* entry
, HttpRequest
* request
)
464 int reason
= refreshCheck(entry
, request
, 10);
465 refreshCounts
[rcHTCP
].total
++;
466 refreshCounts
[rcHTCP
].status
[reason
]++;
467 return (reason
< 200) ? 0 : 1;
472 #if USE_CACHE_DIGESTS
474 refreshCheckDigest(const StoreEntry
* entry
, time_t delta
)
476 int reason
= refreshCheck(entry
,
477 entry
->mem_obj
? entry
->mem_obj
->request
: NULL
,
479 refreshCounts
[rcCDigest
].total
++;
480 refreshCounts
[rcCDigest
].status
[reason
]++;
481 return (reason
< 200) ? 0 : 1;
487 getMaxAge(const char *url
)
490 debugs(22, 3, "getMaxAge: '" << url
<< "'");
492 if ((R
= refreshLimits(url
)))
495 return REFRESH_DEFAULT_MAX
;
500 refreshCountsStats(StoreEntry
* sentry
, struct RefreshCounts
*rc
)
505 storeAppendPrintf(sentry
, "\n\n%s histogram:\n", rc
->proto
);
506 storeAppendPrintf(sentry
, "Count\t%%Total\tCategory\n");
508 #define refreshCountsStatsEntry(code,desc) { \
509 storeAppendPrintf(sentry, "%6d\t%6.2f\t%s\n", \
510 rc->status[code], xpercent(rc->status[code], tot), desc); \
511 sum += rc->status[code]; \
514 refreshCountsStatsEntry(FRESH_REQUEST_MAX_STALE_ALL
,
515 "Fresh: request max-stale wildcard");
516 refreshCountsStatsEntry(FRESH_REQUEST_MAX_STALE_VALUE
,
517 "Fresh: request max-stale value");
518 refreshCountsStatsEntry(FRESH_EXPIRES
,
519 "Fresh: expires time not reached");
520 refreshCountsStatsEntry(FRESH_LMFACTOR_RULE
,
521 "Fresh: refresh_pattern last-mod factor percentage");
522 refreshCountsStatsEntry(FRESH_MIN_RULE
,
523 "Fresh: refresh_pattern min value");
524 refreshCountsStatsEntry(FRESH_OVERRIDE_EXPIRES
,
525 "Fresh: refresh_pattern override expires");
526 refreshCountsStatsEntry(FRESH_OVERRIDE_LASTMOD
,
527 "Fresh: refresh_pattern override lastmod");
528 refreshCountsStatsEntry(STALE_MUST_REVALIDATE
,
529 "Stale: response has must-revalidate");
530 refreshCountsStatsEntry(STALE_RELOAD_INTO_IMS
,
531 "Stale: changed reload into IMS");
532 refreshCountsStatsEntry(STALE_FORCED_RELOAD
,
533 "Stale: request has no-cache directive");
534 refreshCountsStatsEntry(STALE_EXCEEDS_REQUEST_MAX_AGE_VALUE
,
535 "Stale: age exceeds request max-age value");
536 refreshCountsStatsEntry(STALE_EXPIRES
,
537 "Stale: expires time reached");
538 refreshCountsStatsEntry(STALE_MAX_RULE
,
539 "Stale: refresh_pattern max age rule");
540 refreshCountsStatsEntry(STALE_LMFACTOR_RULE
,
541 "Stale: refresh_pattern last-mod factor percentage");
542 refreshCountsStatsEntry(STALE_DEFAULT
,
543 "Stale: by default");
545 tot
= sum
; /* paranoid: "total" line shows 100% if we forgot nothing */
546 storeAppendPrintf(sentry
, "%6d\t%6.2f\tTOTAL\n",
547 rc
->total
, xpercent(rc
->total
, tot
));
549 storeAppendPrintf(sentry
, "\n");
553 refreshStats(StoreEntry
* sentry
)
558 /* get total usage count */
560 for (i
= 0; i
< rcCount
; ++i
)
561 total
+= refreshCounts
[i
].total
;
563 /* protocol usage histogram */
564 storeAppendPrintf(sentry
, "\nRefreshCheck calls per protocol\n\n");
566 storeAppendPrintf(sentry
, "Protocol\t#Calls\t%%Calls\n");
568 for (i
= 0; i
< rcCount
; ++i
)
569 storeAppendPrintf(sentry
, "%10s\t%6d\t%6.2f\n",
570 refreshCounts
[i
].proto
,
571 refreshCounts
[i
].total
,
572 xpercent(refreshCounts
[i
].total
, total
));
574 /* per protocol histograms */
575 storeAppendPrintf(sentry
, "\n\nRefreshCheck histograms for various protocols\n");
577 for (i
= 0; i
< rcCount
; ++i
)
578 refreshCountsStats(sentry
, &refreshCounts
[i
]);
582 refreshRegisterWithCacheManager(void)
584 CacheManager::GetInstance()->
585 registerAction("refresh", "Refresh Algorithm Statistics", refreshStats
, 0, 1);
591 memset(refreshCounts
, 0, sizeof(refreshCounts
));
592 refreshCounts
[rcHTTP
].proto
= "HTTP";
593 refreshCounts
[rcICP
].proto
= "ICP";
596 refreshCounts
[rcHTCP
].proto
= "HTCP";
599 refreshCounts
[rcStore
].proto
= "On Store";
600 #if USE_CACHE_DIGESTS
602 refreshCounts
[rcCDigest
].proto
= "Cache Digests";
605 memset(&DefaultRefresh
, '\0', sizeof(DefaultRefresh
));
606 DefaultRefresh
.pattern
= "<none>";
607 DefaultRefresh
.min
= REFRESH_DEFAULT_MIN
;
608 DefaultRefresh
.pct
= REFRESH_DEFAULT_PCT
;
609 DefaultRefresh
.max
= REFRESH_DEFAULT_MAX
;
611 refreshRegisterWithCacheManager();