--- name: "expire" section: "8" platform: "linux" package: "inn" version: "2.7.4-4" release: "Arch Linux 2026-11-01" platforms: ["linux"] url: "https://new.carta.tech/expire/8" --- # expire(8) Usenet article and history expiration program ## NAME expire - Usenet article and history expiration program ## SYNOPSIS **expire** \[**-iNnptx**\] \[**-d** *dir*\] \[**-f** *file*\] \[**-g** *file*\] \[**-h** *file*\] \[**-r** *reason*\] \[**-s** *size*\] \[**-v** *level*\] \[**-w** *number*\] \[**-z** *file*\] \[*expire.ctl*\] ## DESCRIPTION **expire** scans the [history(5)](/history/5.md)-format text file *pathdb*/history and uses the information recorded in it to purge itself of old news articles. Its behaviour depends on the setting of *groupbaseexpiry* in *inn.conf*. When *groupbaseexpiry* is false, article expiration is primarily done by **expire** based on the expiration rules in *expire.ctl* that match the storage class of each article. The articles and the history entries are removed by **expire**, and then **expireover** does the additional cleanup of removing the overview database entries. History entries of expired articles are removed only if they are older than the number of days specified in the `"/remember/"` line in *expire.ctl*. Articles stored using a storage method that has self-expire functionality like CNFS are by default not affected by **expire**'s primary behaviour (but see the **-N** flag to disable this). When *groupbaseexpiry* is true, article expiration is primarily done by **expireover** based on the expiration rules in *expire.ctl* that match each newsgroup. Articles are removed from the news spool by **expireover**, and then **expire** does some additional cleanup to remove old history database entries. For articles in self-expiring storage methods when *groupbaseexpiry* is set to false in *inn.conf* and the **-N** flag is not given, or for *all* articles when *groupbaseexpiry* is set to true, *expire.ctl* is ignored except the `"/remember/"` line; **expire** then only probes to see if the article still exists, and purges the relevant history entries if appropriate. Regardless the setting of *groupbaseexpiry*, **expireover** should be run along with **expire**, usually via **news.daily** out of cron. Note that **expire** never purges articles which do not match any entry in *expire.ctl*. Also note that if *groupbaseexpiry* is true, the server needs an overview database in order to expire articles in storage backends that are not self-expiring. If you do not plan to have an overview database, it would then be better to only use self-expiring backends like CNFS, as the history entries will still get cleaned up by **expire** when it detects that an article no longer exists in that backend. ## OPTIONS - **-d** *dir* If the **-d** flag is used, then the new *history* file and database is created in the specified directory *dir*. This is useful when the filesystem does not have sufficient space to hold both the old and new history files. When this flag is used, **expire** leaves the server paused and creates a zero-length file named after the new history file, with an extension of `".done"` to indicate that it has successfully completed the expiration. The calling script should install the new history file and unpause the server. The **-r** flag should be used with this flag. - **-f** *file* To specify an alternate history file, use the **-f** flag. This flag is valid when used with the **-d** flag, and the output will be written to the specified file. The default without **-f** is `"history"`. - **-g** *file* If the **-g** flag is given, then a one-line summary equivalent to the output of **-v 1**, except preceded by the current time, will be appended to the specified *file*. - **-h** *file* To specify an alternate input text history file, use the **-h** flag. **expire** uses the old *dbz* database to determine the size of the new one. (If the **-d** flag is not used, the output filename will be the same as the input filename with an extension of `".n"`.) The default without the **-h** flag is *pathdb*/history. - **-i** To ignore the old database, use the **-i** flag. - **-N** The control file is normally ignored for articles in storage methods which have self-expire functionality. If the **-N** flag is used, **expire** still uses the control file for these articles. This parameter is only useful when *groupbaseexpiry* is set to false in *inn.conf*. - **-n** If **innd** is not running, use the **-n** flag and **expire** will not send the `"pause"` or `"go"` commands. (For more details on the commands, see [ctlinnd(8)](/ctlinnd/8.md)). Note that **expire** only needs exclusive access for a very short time -- long enough to see if any new articles arrived since it first hit the end of the file, and to rename the new files to the working files. - **-p** **expire** makes its decisions on the time the article arrived, as found in the *history* file. This means articles are often kept a little longer than with other expiration programs that base their decisions on the article's posting date. To use the article's posting date, use the **-p** flag. - **-r** *reason* **expire** normally sends a `"pause"` command to the local **innd** daemon when it needs exclusive access to the *history* file, using the string `"Expiring"` as the reason. To give a different reason, use the **-r** flag. The process ID will be appended to the reason. When **expire** is finished and the new *history* file is ready, it sends a `"go"` command. See also the **-n** flag. - **-s** *size* **expire** determines the optimal size of the new *history* file from the size of the old one. In case you want to force a specific size, use this flag to optimize the new history database for approximately *size* key-value pairs (i.e. lines in *history*). Accurately specifying the size will create a more efficient database. (The size should be the estimated eventual number of articles, typically the size of the old *history* file, in lines.) - **-t** If the **-t** flag is used, then **expire** will generate a list of the tokens that should be removed on its standard output, and the new *history* file will be left in *history.n*, *history.n.dir*, *history.n.index* and *history.n.hash*. This flag is useful for debugging when used with the **-n** flag. Note that if the **-f** flag is used, then the name specified with that flag will be used instead of *history*. - **-v** *level* The **-v** flag is used to increase the verbosity of the program, generating messages to standard output. The *level* should be a number, where higher numbers result in more output. Level one will print totals of the various actions done (not valid if a new *history* file is not written), level two will print a report on each individual file, while level five results in multiple lines of output for every history line processed. - **-w** *number* Use the **-w** flag to "warp" time so that **expire** thinks it is running at some time other than the current time. The value should be a signed floating point number indicating the number of days to use as the offset. - **-x** If the **-x** flag is used, then **expire** will not create any new history files. This is most useful when combined with the **-n** and **-t** flags to see how different expiration policies would change the amount of disk space used. - **-z** *file* If the **-z** flag is used, then articles are not removed, but their names are appended to the specified *file*. See the description of **delayrm** in [news.daily(8)](/news.daily/8.md). If a filename is specified, it is taken as the control file and parsed according to the rules in *expire.ctl*. A single dash (`"-"`) may be used to read the file from standard input. If no file is specified, the file *pathetc*/expire.ctl is read. ## TOMBSTONE LOG When *expiretombstone* and *groupbaseexpiry* are enabled in *inn.conf*, **expire** consumes the per-cycle deletion log produced by **expireover**, **innd**, and **sm -r** (*expireover.tombstone* and *cancels.tombstone* log files in *pathdb*) so it can drop history entries for those articles without doing a per-article `"SMretrieve(RETR_STAT)"` syscall. An empty tombstone is treated as "no cancels this cycle" and the slow scan is skipped entirely. See [inn.conf(5)](/inn.conf/5.md) under *expiretombstone* for the file lifecycle, locking model, and recovery story. ## HISTORY Written by Rich $alz \ for InterNetNews. Converted to POD by Julien Elie. ## SEE ALSO [ctlinnd(8)](/ctlinnd/8.md), [expire.ctl(5)](/expire.ctl/5.md), [expireover(8)](/expireover/8.md), [history(5)](/history/5.md), [inn.conf(5)](/inn.conf/5.md), [innd(8)](/innd/8.md), [libinn\_dbz(3)](/libinn_dbz/3.md), [libinn\_inndcomm(3)](/libinn_inndcomm/3.md), [news.daily(8)](/news.daily/8.md).