If a message is older than the "first failed" time when computing a

[exim.git] / src / src / retry.c

diff --git a/src/src/retry.c b/src/src/retry.c
index 4566da2e46197af04aab669845f24552f22e8a30..eb4cd46cc18900189a3b0d20fe10295615ef037f 100644 (file)
--- a/src/src/retry.c
+++ b/src/src/retry.c
@@ -1,10 +1,10 @@
-/* $Cambridge: exim/src/src/retry.c,v 1.1 2004/10/07 10:39:01 ph10 Exp $ */
+/* $Cambridge: exim/src/src/retry.c,v 1.7 2006/02/09 14:50:58 ph10 Exp $ */

 
 /*************************************************
 *     Exim - an Internet mail transport agent    *
 *************************************************/
 
 
 /*************************************************
 *     Exim - an Internet mail transport agent    *
 *************************************************/
 

-/* Copyright (c) University of Cambridge 1995 - 2004 */
+/* Copyright (c) University of Cambridge 1995 - 2006 */

 /* See the file NOTICE for conditions of use and distribution. */
 
 /* Functions concerned with retrying unsuccessful deliveries. */
 /* See the file NOTICE for conditions of use and distribution. */
 
 /* Functions concerned with retrying unsuccessful deliveries. */


@@ -46,8 +46,17 @@ if (retry != NULL && retry->rules != NULL)
   for (last_rule = retry->rules;
        last_rule->next != NULL;
        last_rule = last_rule->next);
   for (last_rule = retry->rules;
        last_rule->next != NULL;
        last_rule = last_rule->next);

+  DEBUG(D_transport|D_retry)
+    debug_printf("now=%d received_time=%d diff=%d timeout=%d\n",
+      (int)now, received_time, (int)(now - received_time),
+      last_rule->timeout);

   address_timeout = (now - received_time > last_rule->timeout);
   }
   address_timeout = (now - received_time > last_rule->timeout);
   }

+else
+  {
+  DEBUG(D_transport|D_retry)
+    debug_printf("no retry rule found: assume timed out\n");
+  }

 
 return address_timeout;
 }
 
 return address_timeout;
 }


@@ -340,23 +349,38 @@ retry_config *
 retry_find_config(uschar *key, uschar *alternate, int basic_errno,
   int more_errno)
 {
 retry_find_config(uschar *key, uschar *alternate, int basic_errno,
   int more_errno)
 {

-int replace;
+int replace = 0;

 uschar *use_key, *use_alternate;
 uschar *colon = Ustrchr(key, ':');
 retry_config *yield;
 
 uschar *use_key, *use_alternate;
 uschar *colon = Ustrchr(key, ':');
 retry_config *yield;
 

-/* If there's a colon in the key, temporarily replace it with
-a zero to terminate the string there. */
+/* If there's a colon in the key, there are two possibilities:
+
+(1) This is a key for a host, ip address, and possibly port, in the format
+
+      hostname:ip+port
+
+    In this case, we temporarily replace the colon with a zero, to terminate
+    the string after the host name.
+
+(2) This is a key for a pipe, file, or autoreply delivery, in the format
+
+      pipe-or-file-or-auto:x@y
+
+    where x@y is the original address that provoked the delivery. The pipe or
+    file or auto will start with | or / or >, whereas a host name will start
+    with a letter or a digit. In this case we want to use the original address
+    to search for a retry rule. */

 
 if (colon != NULL)
   {
 
 if (colon != NULL)
   {

-  replace = ':';
-  }
-else
-  {
-  colon = key + Ustrlen(key);
-  replace = 0;
+  if (isalnum(*key))
+    replace = ':';
+  else
+    key = Ustrrchr(key, ':') + 1;   /* Take from the last colon */

   }
   }

+
+if (replace == 0) colon = key + Ustrlen(key);

 *colon = 0;
 
 /* Sort out the keys */
 *colon = 0;
 
 /* Sort out the keys */


@@ -619,10 +643,12 @@ for (i = 0; i < 3; i++)
         DEBUG(D_retry)
           {
           if ((rti->flags & rf_host) != 0)
         DEBUG(D_retry)
           {
           if ((rti->flags & rf_host) != 0)

-            debug_printf("retry for %s (%s) = %s\n", rti->key,
-              addr->domain, retry->pattern);
+            debug_printf("retry for %s (%s) = %s %d %d\n", rti->key,
+              addr->domain, retry->pattern, retry->basic_errno,
+              retry->more_errno);

           else
           else

-            debug_printf("retry for %s = %s\n", rti->key, retry->pattern);
+            debug_printf("retry for %s = %s %d %d\n", rti->key, retry->pattern,
+              retry->basic_errno, retry->more_errno);

           }
 
         /* Set up the message for the database retry record. Because DBM
           }
 
         /* Set up the message for the database retry record. Because DBM


@@ -658,6 +684,16 @@ for (i = 0; i < 3; i++)
         /* Compute how long this destination has been failing */
 
         failing_interval = now - retry_record->first_failed;
         /* Compute how long this destination has been failing */
 
         failing_interval = now - retry_record->first_failed;

+        DEBUG(D_retry) debug_printf("failing_interval=%d message_age=%d\n",
+          failing_interval, message_age);
+
+        /* If the message has been on the queue longer than the recorded time
+        of failure, use the message's age instead. This can happen when some
+        messages can be delivered and others cannot; a successful delivery will
+        reset the first_failed time, and this can lead to a failing message
+        being retried too often. */
+
+        if (message_age > failing_interval) failing_interval = message_age;

 
         /* Search for the current retry rule. The cutoff time of the
         last rule is handled differently to the others. The rule continues
 
         /* Search for the current retry rule. The cutoff time of the
         last rule is handled differently to the others. The rule continues


@@ -712,7 +748,14 @@ for (i = 0; i < 3; i++)
 
         This implements "timeout this rule if EITHER the host (or routing or
         directing) has been failing for more than the maximum time, OR if the
 
         This implements "timeout this rule if EITHER the host (or routing or
         directing) has been failing for more than the maximum time, OR if the

-        message has been on the queue for more than the maximum time." */
+        message has been on the queue for more than the maximum time."
+
+        February 2006: It is possible that this code is no longer needed
+        following the change to the retry calculation to use the message age if
+        it is larger than the time since first failure. It may be that the
+        expired flag is always set when the other conditions are met. However,
+        this is a small bit of code, and it does no harm to leave it in place,
+        just in case. */

 
         if (received_time <= retry_record->first_failed &&
             addr == endaddr && !retry_record->expired && rule != NULL)
 
         if (received_time <= retry_record->first_failed &&
             addr == endaddr && !retry_record->expired && rule != NULL)


@@ -739,15 +782,24 @@ for (i = 0; i < 3; i++)
         if (rule == NULL) next_try = now; else
           {
           if (rule->rule == 'F') next_try = now + rule->p1;
         if (rule == NULL) next_try = now; else
           {
           if (rule->rule == 'F') next_try = now + rule->p1;

-          else  /* assume rule = 'G' */
+          else  /* rule = 'G' or 'H' */

             {
             int last_predicted_gap =
               retry_record->next_try - retry_record->last_try;
             int last_actual_gap = now - retry_record->last_try;
             int lastgap = (last_predicted_gap < last_actual_gap)?
               last_predicted_gap : last_actual_gap;
             {
             int last_predicted_gap =
               retry_record->next_try - retry_record->last_try;
             int last_actual_gap = now - retry_record->last_try;
             int lastgap = (last_predicted_gap < last_actual_gap)?
               last_predicted_gap : last_actual_gap;

-            next_try = now + ((lastgap < rule->p1)? rule->p1 :
-               (lastgap * rule->p2)/1000);
+            int next_gap = (lastgap * rule->p2)/1000;
+            if (rule->rule == 'G')
+              {
+              next_try = now + ((lastgap < rule->p1)? rule->p1 : next_gap);
+              }
+            else  /* The 'H' rule */
+              {
+              next_try = now + rule->p1;
+              if (next_gap > rule->p1)
+                next_try += random_number(next_gap - rule->p1);
+              }

             }
           }
 
             }
           }
 


@@ -842,6 +894,9 @@ for (i = 0; i < 3; i++)
           setflag(addr, af_retry_timedout);
           addr->message = (addr->message == NULL)? US"retry timeout exceeded" :
             string_sprintf("%s: retry timeout exceeded", addr->message);
           setflag(addr, af_retry_timedout);
           addr->message = (addr->message == NULL)? US"retry timeout exceeded" :
             string_sprintf("%s: retry timeout exceeded", addr->message);

+          addr->user_message = (addr->user_message == NULL)?
+            US"retry timeout exceeded" :
+            string_sprintf("%s: retry timeout exceeded", addr->user_message);

           log_write(0, LOG_MAIN, "** %s%s%s%s: retry timeout exceeded",
             addr->address,
            (addr->parent == NULL)? US"" : US" <",
           log_write(0, LOG_MAIN, "** %s%s%s%s: retry timeout exceeded",
             addr->address,
            (addr->parent == NULL)? US"" : US" <",