1BUFFINDEXED.CONF(5)       InterNetNews Documentation       BUFFINDEXED.CONF(5)
2
3
4

NAME

6       buffindexed.conf - Configuration for the buffindexed overview method
7

DESCRIPTION

9       buffindexed.conf, found in pathetc, specifies the buffers that the
10       buffindexed overview method should use.  It is required if the server
11       uses buffindexed (as configured by the ovmethod parameter in inn.conf).
12
13       Buffindexed uses pre-built buffer files to store overview data and
14       indexes to that data.  The buffers are divided into 8 KB internally,
15       and a given block is used either for overview data or for index data.
16       A block is always allocated to a single newsgroup and is never shared
17       among newsgroups.
18
19       In addition to the buffers, buffindexed also stores information in a
20       file named group.index in pathdb.  (This file should not be mistaken
21       for the one named group.index in pathoverview which is used by the
22       tradindexed overview method.)  It contains information about each
23       newsgroup:  the pointer to the index block for the newsgroup, the high
24       mark, the low mark, the flag of the group, the number of articles, and
25       so forth.  This file is created automatically when all buffers are
26       initialized and should not be manually edited.
27
28       Buffindexed buffers are of fixed size, so buffindexed will never use
29       more space than what is available in those buffers.  If all buffers are
30       full, innd will throttle when it attempts to store overview information
31       for any additional articles until space is freed (with expireover, for
32       instance) or another buffer is added.  This is unlike the CNFS storage
33       method.
34
35       You can see the current usage of the buffers with the -o option to
36       inndf.
37
38       In the buffindexed.conf file, blank lines and lines beginning with a
39       number sign ("#") are ignored.  All other lines must be of the format:
40
41           <index>:<filename>:<size>
42
43       The order of lines is not significant.
44
45       <index> is the index of this overview buffer and must be unique.  Other
46       than that constraint, it can be any number between 0 and 65535.
47
48       <filename> is the path to the buffer.  The length of the path should
49       not be longer than 63 characters.
50
51       <size> is the length of the buffer in kilobytes (1 KB = 1024 bytes).
52       If <filename> does not specify a special device, the file size of the
53       buffer must be <size> * 1024 bytes.  If it does specify a special
54       device, that device must have at least <size> space available.  For
55       more information on setting up the buffers, see "CREATING BUFFERS".
56
57       An example of buffindexed.conf file can be:
58
59           0:<pathoverview in inn.conf>/OV1:1536000
60           1:<pathoverview in inn.conf>/OV2:1536000
61
62       When you first start innd with everything configured properly, you
63       should see messages like this in pathlog/news.notice:
64
65           Aug 27 00:00:00 kevlar innd: buffindexed: no magic cookie found
66               for ovbuff 0, initializing
67
68       You MUST recreate overview completely using makehistory if you remove
69       or replace buffers.  However, new buffers can be added without any
70       special care (other than restarting innd after modifying
71       buffindexed.conf).  If you need to rebuild overview, you should zero
72       all of the buffers first.
73
74       We recommend not to reserve too much spare space in existing buffers,
75       so that to minimize the duration of the expireover process, and to just
76       add new buffers when space left is low (see the result of "inndf -no").
77       Plan on needing at least 0.65 KB for every article in your spool (not
78       counting crossposts).  So, if you have 5 million articles, you'll need
79       at least 3.25 GB of disk space for buffindexed.
80

CREATING BUFFERS

82       There are two methods to create a new buffindexed buffer:
83
84       1.  Create a large file on top of a regular file system.  The easiest
85           way to do this is probably with dd(1), using a command like:
86
87               dd if=/dev/zero of=/path/to/cycbuff bs=1024 count=<size>
88
89           where <size> is the size from the relevant line in
90           buffindexed.conf.
91
92           This is the simplest method, but has the disadvantage that very
93           large files on regular file systems can be fairly slow to access,
94           particularly at the end of the file, and INN incurs unnecessary
95           file system overhead when accessing the buffer.
96
97       2.  Use block devices directly.  If your operating system allows you to
98           call mmap() on block devices (Solaris and recent versions of Linux
99           do, FreeBSD at last report does not), this method can avoid all of
100           the native file system overhead.  Note, however, that Solaris has
101           problems with byte range locking on block devices, and therefore
102           this method should not be used on Solaris.
103
104           Partition the disk.  If you're using Solaris, set up your
105           partitions to avoid the first cylinder of the disk (or otherwise
106           the buffindexed header will overwrite the disk partition table and
107           render the buffers inaccessible).  Then, create device files for
108           each block device you're going to use.
109
110           It's not recommended to use the block device files in /dev, since
111           the news system doesn't have permission to write to them and
112           changing the permissions of the system device files may affect
113           something else.  Instead, use mknod(1) to create a new set of block
114           devices (in somewhere like pathspool/overview that's only writable
115           by the news user).  To do this, run "ls -Ll" on the devices in /dev
116           that correspond to the block devices that you want to use.  The
117           major and minor device numbers are in the fifth and sixth columns
118           (right before the date), respectively.  Then run mknod like:
119
120               mknod <filename> b <major> <minor>
121
122           where <filename> is the path to the device to create (matching the
123           <filename> part of the buffindexed configuration line) and <major>
124           and <minor> are the major and minor device numbers as discovered
125           above.
126
127           Here's a short script to do this when given the path to the system
128           device file as an argument:
129
130               #!/bin/sh
131               base=`echo "$1" | sed 's%.*/%%'`
132               major=`ls -Ll "$1" | awk '{print $5}' | tr -d ,`
133               minor=`ls -Ll "$1" | awk '{print $6}`
134               mkdir -p <pathoverview in inn.conf>
135               mknod <pathoverview>/"$base" b "$major" "$minor"
136               chown news:news <pathoverview>/"$base"
137               chmod 644 <pathoverview>/"$base"
138
139           Make sure that the created files are owned by the news user and
140           news group, as specified at configure time (the default being
141           "news" for both).  Also make sure that the permissions on the
142           devices allow the news user to read and write.
143

HISTORY

145       Written by Katsuhiro Kondou <kondou@nec.co.jp> for InterNetNews.
146       Converted to POD by Russ Allbery <eagle@eyrie.org>.
147
148       $Id: buffindexed.conf.pod 10525 2021-01-20 11:51:15Z iulius $
149

SEE ALSO

151       expireover(8), inn.conf(5), inndf(8), makehistory(8).
152
153
154
155INN 2.6.4                         2021-01-21               BUFFINDEXED.CONF(5)
Impressum