Ovdb is a storage method that uses the BerkeleyDB library to store overview
data. It requires version 2.6.x or later of the BerkeleyDB library; it is
known to work well with version 2.7.7. There is also a 3.0.55, but it is
not very stable (though ovdb will compile correctly with it). Check the
web page http://www.avalon.net/~hakehoe/inn/ for the latest on BerkeleyDB
compatibility.
Ovdb makes use of the full transaction/logging/locking functionality of the
BerkeleyDB environment. BerkeleyDB may be downloaded from
http://www.sleepycat.com .
The BerkeleyDB distribution is (by default) installed into
/usr/local/BerkeleyDB, with subdirectories lib/, include/, and
bin/ for the library, headers, and support tools, respectively.
Version 3.x uses /usr/local/BerkeleyDB.3.x instead.
Ovdb uses some of the support tools in the bin/ subdirectory; namely
db_archive, db_checkpoint, and db_deadlock.
INSTALLATION
To build ovdb support into INN, specify the option `--with-berkeleydb'
when running the configure script. By default, configure will search
for a BerkeleyDB tree in several likely locations, and choose the highest
version (based on the name of the directory, e.g., BerkeleyDB.3.0)
that it finds. There will be a message in the configure output
indicating the chosen pathname.
You can override this pathname by adding a path to the option, e.g.,
`--with-berkeleydb=/usr/BerkeleyDB.3.1'. This directory is expected
to have subdirectories include, lib, and bin containing
db.h, the library itself, and the db_* utilities, respectively.
The ovdb database will take up more disk space for a given spool than
the other overview methods. Plan on needing at least 1.1 KB for every
article in your spool (not counting crossposts). So, if you have 5
million articles, you'll need at least 5.5 GB of disk space for ovdb.
With BerkeleyDB 2.x, the db files are 'grow only'; the library will
not shrink them, even if data is removed. So, reserving extra space
above the estimate is a good idea.
Plus, you'll need additional space for transaction logs: at least 40 MB,
preferably 100 MB or more. (The logs may be on a different filesystem
-- see the DB_CONFIG section.)
CONFIGURATION
To enable ovdb, set the ovmethod parameter in inn.conf to `ovdb'.
The ovdb database is stored in the directory specified by the pathoverview
paramter in inn.conf. This is the ``DB_HOME'' directory. To start out,
this directory should be empty (other than an optional DB_CONFIG file,
see the DB_CONFIG section for details) and innd (or makehistory) will
create the files as necessary in that directory.
Other parameters for configuring ovdb are in the optional ovdb.conf
configuration file. See also the sample ovdb.conf.
numdbfiles
Overview data is split between this many files. Currently,
innd will keep all of the files open, so don't set this too high
or innd may run out of file descriptors. The nnrpds only open one
at a time, regardless. May be set to one, or just a few, but only
do that if your OS supports large (>2G) files.
Warning: do NOT change this on an already established database.
Default is 32.
txn_nosync
If txn_nosync is set to false, BerkeleyDB flushes the log after every
transaction. This minimizes the number of transactions that may be lost
in the event of a crash, but results in significantly degraded
performance. Default is true.
cachesize
Size of the memory pool cache, in Kilobytes. The cache will have a
backing store file in the DB directory which will be at least as big.
In general, the bigger the cache, the better. Use `db_stat -m' to see
cache hit percentages. If they're less than 80%, try increasing the
cache size. To make a change of this parameter take effect with an
existing database, remove the __db* files while there are no
database processes running. Default is 8000 K, which may be
adequate for small to medium-sized servers. Full-feed servers will
probably need at least 14000.
pagesize
Sets the page size for the DB files (in bytes). Must be a power of 2.
Best choices are 4096, 8192, or 16384. The default is 8192.
You can not change the pagesize of an existing DB, it must be recreated.
minkey
Sets the minimum number of keys per page. See the BerkeleyDB
documentation for more info. Default is based on page size:
default_minkey = pagesize / 2048
You can not change the minkey of an existing DB, it must be recreated.
DB_CONFIG
A file called DB_CONFIG may be placed in the database directory to
customize where the various database files and transaction logs are
written. By default, all of the files are written in the ``DB_HOME''
directory. One way to improve performance is to put the transaction
logs on a different disk. To do this, put:
DB_LOG_DIR /path/to/logs
in the DB_CONFIG file. If the pathname you give starts with a /,
it is treated as an absolute path; otherwise, it is relative to the
``DB_HOME'' directory. Make sure that any directories you specify
exist and have proper ownership/mode before starting INN, because
they won't be created automatically. Also, don't change the DB_CONFIG
file while anything that uses ovdb is running.
The DB_CONFIG functionality is part of BerkeleyDB itself.
RUNNING
You don't need to do anything special when starting or stopping INN
when using ovdb, as long as you use rc.news. When starting,
rc.news:
*
Runs `ovdb_recover'. This will perform any needed recovery on the
database; e.g., if there was a crash that may have left the database
in an inconsistent state. This is done before innd is started. If
the innd.pid file is present but innd is not running (i.e., unclean
shutdown), rc.news will supply the `-f' argument to `ovdb_recover'
*
Runs `dbprocs start'. This starts the DB housekeeping processes.
And when stopping INN, rc.news calls `dbprocs stop' after the
other INN processes have been shut down.
DIAGNOSTICS
Problems relating to ovdb are logged to news.err with ``OVDB'' in the
error message.
If a program accessing the database crashes, or otherwise exits
uncleanly, it might leave a stale lock in the database. This lock
could cause other processes to deadlock on that stale lock. To
fix this, shut down all news processes (using
`kill -9' if necessary) then run `ovdb_recover -f' as the
news user, e.g.: `su news -c "/PATHBIN/ovdb_recover -f"'
This will remove all locks plus repair any damage caused by
killing deadlocked processes.
FILES
inn.conf
The ovmethod and pathoverview parameters are relevant to ovdb.
ovdb.conf
Optional configuration file for tuning. See CONFIGURATION section above.
pathoverview
Directory where the database goes. BerkeleyDB calls it the 'DB_HOME'
directory.
pathoverview/DB_CONFIG
Optional file to configure the layout of the database files.
TO DO
I've modified innshellvars and innshellvars.pl to add the BerkeleyDB
bin dir to PATH and set DB_HOME to innconf->pathoverview. Since I
don't know TCL, someone needs to modify innshellvars.tcl in the same
way.
Implement a way to limit how many databases can be open at once (to
reduce file descriptor usage); maybe using something similar to the
cache code in ov3.c
