INN commit: trunk (doc/pod/inn.conf.pod samples/inn.conf.in)

[INN Commit]

    Date: Sunday, March 15, 2009 @ 09:25:23
  Author: iulius
Revision: 8383

Increase the default settings of overcachesize from 15 to 128
and msgidcachesize from 10000 to 16000.

Modified:
  trunk/doc/pod/inn.conf.pod
  trunk/samples/inn.conf.in

----------------------+
 doc/pod/inn.conf.pod |   16 ++++++++--------
 samples/inn.conf.in  |    4 ++--
 2 files changed, 10 insertions(+), 10 deletions(-)

Modified: doc/pod/inn.conf.pod
===================================================================
--- doc/pod/inn.conf.pod	2009-03-15 16:07:43 UTC (rev 8382)
+++ doc/pod/inn.conf.pod	2009-03-15 16:25:23 UTC (rev 8383)
@@ -424,8 +424,8 @@
 may be able to use the C<limit> command to see how many open file
 descriptors your operating system allows.  innd(8) also uses an open file
 descriptor for each incoming feed and outgoing channel or batch file, and
-if it runs out of open file descriptors it may throttle and stop accepting
-new news.  The default value is C<15> (which is probably way too low if
+if it runs out of open file descriptors, it may throttle and stop accepting
+new news.  The default value is C<128> (which is probably still too low if
 you have a large number of file descriptors available).
 
 This setting is ignored unless I<ovmethod> is set to C<tradindexed>.
@@ -565,12 +565,12 @@
 
 =item I<msgidcachesize>
 
-How many cache slots to reserve for Message ID to storage token
-translations.  When serving overview data to clients (NEWNEWS, XOVER
-etc.), nnrpd(8) can cache the storage token associated with a Message
-ID and save the cost of looking it up in the history file; for some
-configurations setting this parameter can save more than 90% of the
-wall clock time for a session.  The default value is 10000.
+How many cache slots to reserve for message-IDs to storage token
+translations.  When serving overview data to clients (NEWNEWS, OVER,
+etc.), nnrpd(8) can cache the storage token associated with a message-ID
+and save the cost of looking it up in the F<history> file; for some
+configurations, setting this parameter can save more than 90% of the
+wall clock time for a session.  The default value is C<16000>.
 
 =item I<nfsreader>
 

Modified: samples/inn.conf.in
===================================================================
--- samples/inn.conf.in	2009-03-15 16:07:43 UTC (rev 8382)
+++ samples/inn.conf.in	2009-03-15 16:25:23 UTC (rev 8383)
@@ -66,7 +66,7 @@
 groupbaseexpiry:        true
 mergetogroups:          false
 nfswriter:              false
-overcachesize:          15
+overcachesize:          128
 #ovgrouppat:
 storeonxref:            true
 useoverchan:            false
@@ -79,7 +79,7 @@
 articlemmap:            true
 clienttimeout:          1800
 initialtimeout:         10
-msgidcachesize:         10000
+msgidcachesize:         16000
 nfsreader:              false
 nfsreaderdelay:         60
 nnrpdcheckart:          true