src/refresh.cc

   1
   2
   3 /*
   4  * $Id: refresh.cc,v 1.38 1998/10/17 04:41:48 rousskov Exp $
   5  *
   6  * DEBUG: section 22    Refresh Calculation
   7  * AUTHOR: Harvest Derived
   8  *
   9  * SQUID Internet Object Cache  http://squid.nlanr.net/Squid/
  10  * ----------------------------------------------------------
  11  *
  12  *  Squid is the result of efforts by numerous individuals from the
  13  *  Internet community.  Development is led by Duane Wessels of the
  14  *  National Laboratory for Applied Network Research and funded by the
  15  *  National Science Foundation.  Squid is Copyrighted (C) 1998 by
  16  *  Duane Wessels and the University of California San Diego.  Please
  17  *  see the COPYRIGHT file for full details.  Squid incorporates
  18  *  software developed and/or copyrighted by other sources.  Please see
  19  *  the CREDITS file for full details.
  20  *
  21  *  This program is free software; you can redistribute it and/or modify
  22  *  it under the terms of the GNU General Public License as published by
  23  *  the Free Software Foundation; either version 2 of the License, or
  24  *  (at your option) any later version.
  25  *
  26  *  This program is distributed in the hope that it will be useful,
  27  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  28  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  29  *  GNU General Public License for more details.
  30  *
  31  *  You should have received a copy of the GNU General Public License
  32  *  along with this program; if not, write to the Free Software
  33  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  34  *
  35  */
  36
  37 #ifndef USE_POSIX_REGEX
  38 #define USE_POSIX_REGEX         /* put before includes; always use POSIX */
  39 #endif
  40
  41 #include "squid.h"
  42
  43 typedef enum { rcHTTP, rcICP, rcCDigest, rcCount } refreshCountsEnum;
  44
  45 static struct RefreshCounts {
  46     const char *proto;
  47     int total;
  48     int revalidate_stale;
  49     int request_max_age_stale;
  50     int response_expires_stale;
  51     int response_expires_fresh;
  52     int conf_max_age_stale;
  53     int last_modified_factor_fresh;
  54     int last_modified_factor_stale;
  55     int conf_min_age_fresh;
  56     int default_stale;
  57 } refreshCounts[rcCount];
  58
  59 /*
  60  * Defaults:
  61  *      MIN     NONE
  62  *      PCT     20%
  63  *      MAX     3 days
  64  */
  65 #define REFRESH_DEFAULT_MIN     (time_t)0
  66 #define REFRESH_DEFAULT_PCT     0.20
  67 #define REFRESH_DEFAULT_MAX     (time_t)259200
  68
  69 static const refresh_t *refreshLimits(const char *);
  70 static const refresh_t *refreshUncompiledPattern(const char *);
  71 static OBJH refreshStats;
  72
  73 static const refresh_t *
  74 refreshLimits(const char *url)
  75 {
  76     const refresh_t *R;
  77     for (R = Config.Refresh; R; R = R->next) {
  78         if (!regexec(&(R->compiled_pattern), url, 0, 0, 0))
  79             return R;
  80     }
  81     return NULL;
  82 }
  83
  84 static const refresh_t *
  85 refreshUncompiledPattern(const char *pat)
  86 {
  87     const refresh_t *R;
  88     for (R = Config.Refresh; R; R = R->next) {
  89         if (0 == strcmp(R->pattern, pat))
  90             return R;
  91     }
  92     return NULL;
  93 }
  94
  95 /*  return 1 if the entry must be revalidated within delta seconds
  96  *         0 otherwise
  97  *
  98  *  note: request maybe null (e.g. for cache digests build)
  99  */
 100 static int
 101 refreshCheck(const StoreEntry * entry, request_t * request, time_t delta, struct RefreshCounts *rc)
 102 {
 103     const refresh_t *R;
 104     const char *uri = NULL;
 105     time_t min = REFRESH_DEFAULT_MIN;
 106     double pct = REFRESH_DEFAULT_PCT;
 107     time_t max = REFRESH_DEFAULT_MAX;
 108 #if HTTP_VIOLATIONS
 109     int override_expire = 0;
 110     int override_lastmod = 0;
 111     int reload_into_ims = 0;
 112     int ignore_reload = 0;
 113 #endif
 114     const char *pattern = "<none>";
 115     time_t age;
 116     double factor;
 117     time_t check_time = squid_curtime + delta;
 118     if (entry->mem_obj)
 119         uri = entry->mem_obj->url;
 120     else
 121     if (request)
 122         uri = urlCanonical(request);
 123
 124     debug(22, 3) ("refreshCheck(%s): '%s'\n", rc->proto, uri ? uri : "<none>");
 125     rc->total++;
 126     if (EBIT_TEST(entry->flags, ENTRY_REVALIDATE)) {
 127         debug(22, 3) ("refreshCheck: YES: Required Authorization\n");
 128         rc->revalidate_stale++;
 129         return 1;
 130     }
 131     if ((R = uri ? refreshLimits(uri) : refreshUncompiledPattern("."))) {
 132         min = R->min;
 133         pct = R->pct;
 134         max = R->max;
 135         pattern = R->pattern;
 136 #if HTTP_VIOLATIONS
 137         override_expire = R->flags.override_expire;
 138         override_lastmod = R->flags.override_lastmod;
 139         reload_into_ims = R->flags.reload_into_ims;
 140         ignore_reload = R->flags.ignore_reload;
 141 #endif
 142     }
 143 #if HTTP_VIOLATIONS
 144     if (!reload_into_ims)
 145         reload_into_ims = Config.onoff.reload_into_ims;
 146 #endif
 147     debug(22, 3) ("refreshCheck: Matched '%s %d %d%% %d'\n",
 148         pattern, (int) min, (int) (100.0 * pct), (int) max);
 149     age = check_time - entry->timestamp;
 150     debug(22, 3) ("refreshCheck: age = %d\n", (int) age);
 151     debug(22, 3) ("\tcheck_time:\t%s\n", mkrfc1123(check_time));
 152     debug(22, 3) ("\tentry->timestamp:\t%s\n", mkrfc1123(entry->timestamp));
 153     /* request-specific checks */
 154     if (request) {
 155 #if HTTP_VIOLATIONS
 156         if (request->flags.nocache_hack) {
 157             if (ignore_reload) {
 158                 /* The clients no-cache header is ignored */
 159                 debug(22, 3) ("refreshCheck: MAYBE: ignore-reload\n");
 160             } else if (reload_into_ims) {
 161                 /* The clients no-cache header is changed into a IMS query */
 162                 debug(22, 3) ("refreshCheck: YES: reload-into-ims\n");
 163                 return 1;
 164             } else {
 165                 /* The clients no-cache header is not overridden on this request */
 166                 debug(22, 3) ("refreshCheck: YES: client reload\n");
 167                 request->flags.nocache = 1;
 168                 return 1;
 169             }
 170         }
 171 #endif
 172         if (request->max_age > -1) {
 173             if (age > request->max_age) {
 174                 debug(22, 3) ("refreshCheck: YES: age > client-max-age\n");
 175                 rc->request_max_age_stale++;
 176                 return 1;
 177             }
 178         }
 179     }
 180 #if HTTP_VIOLATIONS
 181     if (override_expire && age <= min) {
 182         debug(22, 3) ("refreshCheck: NO: age < min && override_expire\n");
 183         return 0;
 184     }
 185 #endif
 186     if (entry->expires > -1) {
 187         if (entry->expires <= check_time) {
 188             debug(22, 3) ("refreshCheck: YES: expires <= curtime\n");
 189             rc->response_expires_stale++;
 190             return 1;
 191         } else {
 192             debug(22, 3) ("refreshCheck: NO: expires > curtime\n");
 193             rc->response_expires_fresh++;
 194             return 0;
 195         }
 196     }
 197     if (age > max) {
 198         debug(22, 3) ("refreshCheck: YES: age > max\n");
 199         rc->conf_max_age_stale++;
 200         return 1;
 201     }
 202 #if HTTP_VIOLATIONS
 203     if (override_lastmod && age <= min) {
 204         debug(22, 3) ("refreshCheck: NO: age < min && override_lastmod\n");
 205         return 0;
 206     }
 207 #endif
 208     if (entry->lastmod > -1 && entry->timestamp > entry->lastmod) {
 209         factor = (double) age / (double) (entry->timestamp - entry->lastmod);
 210         debug(22, 3) ("refreshCheck: factor = %f\n", factor);
 211         if (factor < pct) {
 212             debug(22, 3) ("refreshCheck: NO: factor < pct\n");
 213             rc->last_modified_factor_fresh++;
 214             return 0;
 215         } else {
 216             debug(22, 3) ("refreshCheck: YES: factor >= pct\n");
 217             rc->last_modified_factor_stale++;
 218             return 1;
 219         }
 220     }
 221     if (age <= min) {
 222         debug(22, 3) ("refreshCheck: NO: age < min\n");
 223         rc->conf_min_age_fresh++;
 224         return 0;
 225     }
 226     debug(22, 3) ("refreshCheck: YES: default stale\n");
 227     rc->default_stale++;
 228     return 1;
 229 }
 230
 231 /* refreshCheck... functions below are protocol-specific wrappers around
 232  * refreshCheck() function above */
 233
 234 int
 235 refreshCheckHTTP(const StoreEntry * entry, request_t * request) {
 236     return refreshCheck(entry, request, 0, refreshCounts + rcHTTP);
 237 }
 238
 239 int
 240 refreshCheckICP(const StoreEntry * entry, request_t * request) {
 241     return refreshCheck(entry, request, 30, refreshCounts + rcICP);
 242 }
 243
 244 int
 245 refreshCheckDigest(const StoreEntry * entry, time_t delta) {
 246     return refreshCheck(entry, NULL, delta, refreshCounts + rcCDigest);
 247 }
 248
 249 time_t
 250 getMaxAge(const char *url)
 251 {
 252     const refresh_t *R;
 253     debug(22, 3) ("getMaxAge: '%s'\n", url);
 254     if ((R = refreshLimits(url)))
 255         return R->max;
 256     else
 257         return REFRESH_DEFAULT_MAX;
 258 }
 259
 260 static void
 261 refreshCountsStats(StoreEntry * sentry, struct RefreshCounts *rc)
 262 {
 263     storeAppendPrintf(sentry, "\n\n%s histogram:\n", rc->proto);
 264     storeAppendPrintf(sentry, "Category\tCount\t%%Total\n");
 265
 266     storeAppendPrintf(sentry, "revalidate_stale\t%6d\t%6.2f\n",
 267         rc->revalidate_stale, xpercent(rc->revalidate_stale, rc->total));
 268     storeAppendPrintf(sentry, "request_max_age_stale\t%6d\t%6.2f\n",
 269         rc->request_max_age_stale, xpercent(rc->request_max_age_stale, rc->total));
 270     storeAppendPrintf(sentry, "response_expires_stale\t%6d\t%6.2f\n",
 271         rc->response_expires_stale, xpercent(rc->response_expires_stale, rc->total));
 272     storeAppendPrintf(sentry, "response_expires_fresh\t%6d\t%6.2f\n",
 273         rc->response_expires_fresh, xpercent(rc->response_expires_fresh, rc->total));
 274     storeAppendPrintf(sentry, "conf_max_age_stale\t%6d\t%6.2f\n",
 275         rc->conf_max_age_stale, xpercent(rc->conf_max_age_stale, rc->total));
 276     storeAppendPrintf(sentry, "last_modified_factor_fresh\t%6d\t%6.2f\n",
 277         rc->last_modified_factor_fresh, xpercent(rc->last_modified_factor_fresh, rc->total));
 278     storeAppendPrintf(sentry, "last_modified_factor_stale\t%6d\t%6.2f\n",
 279         rc->last_modified_factor_stale, xpercent(rc->last_modified_factor_stale, rc->total));
 280     storeAppendPrintf(sentry, "conf_min_age_fresh\t%6d\t%6.2f\n",
 281         rc->conf_min_age_fresh, xpercent(rc->conf_min_age_fresh, rc->total));
 282     storeAppendPrintf(sentry, "default_stale\t%6d\t%6.2f\n",
 283         rc->default_stale, xpercent(rc->default_stale, rc->total));
 284     storeAppendPrintf(sentry, "total\t%6d\t%6.2f\n",
 285         rc->total, xpercent(rc->total, rc->total));
 286 }
 287
 288 static void
 289 refreshStats(StoreEntry * sentry)
 290 {
 291     int i;
 292     int total = 0;
 293
 294     /* get total usage count */
 295     for (i = 0; i < rcCount; ++i)
 296         total += refreshCounts[i].total;
 297
 298     /* protocol usage histogram */
 299     storeAppendPrintf(sentry, "\nRefreshCheck calls per protocol\n\n");
 300     storeAppendPrintf(sentry, "Protocol\t#Calls\t%%Calls\n");
 301     for (i = 0; i < rcCount; ++i)
 302         storeAppendPrintf(sentry, "%10s\t%6d\t%6.2f\n",
 303             refreshCounts[i].proto,
 304             refreshCounts[i].total,
 305             xpercent(refreshCounts[i].total, total));
 306
 307     /* per protocol histograms */
 308     storeAppendPrintf(sentry, "\n\nRefreshCheck histograms for various protocols\n");
 309     for (i = 0; i < rcCount; ++i)
 310         refreshCountsStats(sentry, refreshCounts + i);
 311 }
 312
 313 void
 314 refreshInit()
 315 {
 316     memset(refreshCounts, 0, sizeof(refreshCounts));
 317     refreshCounts[rcHTTP].proto = "HTTP";
 318     refreshCounts[rcICP].proto = "ICP";
 319     refreshCounts[rcCDigest].proto = "Cache Digests";
 320
 321     cachemgrRegister("refresh",
 322         "Refresh Algorithm Statistics",
 323         refreshStats,
 324         0,
 325         1);
 326 }