QuickWeed - a fast but conservative Vesta-2 weeder
QuickWeed [-n] [-i] [-p] [-no-cache] [-cf config-file] ...
QuickWeed is a fast but conservative weeder for Vesta-2. As opposed to the real weeder VestaWeed(1), QuickWeed treats every entry in the cache as a root of its mark phase. Hence, every cache entry is retained. The only files deleted are those that are not mentioned as deriveds in any cache entry and not mentioned as sources in any directory in the repository.
Normally, QuickWeed cannot be run if there is an incomplete VestaWeed in progress. In particular, if VestaWeed (or the cache) crashes before it has finished, you must run VestaWeed again to finish the old weed and clean up its state before a new weed can be started by either QuickWeed or VestaWeed. QuickWeed will exit with an informative message if it detects that another VestaWeed or QuickWeed is in progress.
QuickWeed also has a special mode for use with a repository that has no associated cache (e.g. one used for replication but never for building). If both -i and -no-cache are specified, QuickWeed will cause the repository to do its part of the weeding process but keeps no derived files. This should only be used if there is no cache associated with the repository, as it invalidates any such cache.
Since QuickWeed must write a file into the Vesta repository specifying which deriveds to keep, it must be run as the Vesta administrator. This requirement does not apply if the -n option is being used.
- If the -n option is specified, QuickWeed computes the short-ID's of the derived files to be kept and writes them to the standard output, but it does not actually do the work of weeding any source or derived files. Also implies the -p option.
- Suppresses checkpointing the repository. By default, QuickWeed tells the repository to checkpoint itself after the weed is complete, which compacts the repository's memory and speeds up recovery the next time the repository is restarted.
- -cf config-file
- The -cf option can be specified any number of times. Each instance names a new configuration file to be consulted. If config-file is a pathname ending in "/", the filename "vesta.cfg" is automatically appended. For each additional configuration file consulted, QuickWeed contacts the cache server named in the configuration file and records the names of the derived files in that cache server's graph log. Thus, all the derived files in all the named caches are protected from deletion.
Specifying more than one cache to QuickWeed is useful only in the very unusual situation that multiple caches are being used with the same repository. This is not ordinarily done, and in fact, VestaWeed does not support it. QuickWeed supports multiple caches only for historical reasons; the feature was useful when the cache server was first being implemented and debugged.
- Ignores the default configuration file. By default, QuickWeed determines the deriveds to retain by contacting the cache named in the user's Vesta configuration file. When -i is specified, the user's configuration file is ignored. The program reports an error if -i is specified without -cf, unless -no-cache is also sepcified.
- Allows all caches to be ignored completely. In other words, it enables performing the repository's part of the weeding process but keeps no derived files. This is intended for use with repositories that do not have associated caches (e.g. ones used for replication but never for building). Both -i and -no-cache must be specified together to activate this feature. The program reports an error if -no-cache is specified without -i or with -cf.
QuickWeed prints diagnostic messages to the standard error output. It prints the start and end times of each graph log scan, the derived weed, and the deletion phase. (The repository performs source weeding together with the deletion phase; see its event log for the messages it prints.) QuickWeed also prints the names of the configuration files consulted, the host and port of the corresponding cache server, the number of graph log entries processed, and the total number of derived files named in those entries. (The latter figure may double-count if the same derived file is named in more than one graph log entry).
- $ QuickWeed -n > /tmp/deriveds
- Runs QuickWeed, but doesn't do any deletions. The list of deriveds to keep is saved in the named file. The following output appears on the standard error output:09:48:13.156 -- Started graph log scan Configuration file: /tmp_mnt/shellgame/r/dlusers9/heydon/vesta.cfg Hostname:port: rafael.pa.dec.com:21763 Graph log: /rafael/vesta-srv/cache/graphLog 09:48:13.460 -- Finished graph log scan Graph log entries processed = 1114 Derived files found = 2293
- $ QuickWeed
- Same as above, only this time, the source and derived weeds do occur. Here is the output from this run:Writing ShortIds of deriveds to keep to file fcb0c000 14:32:41.935 -- Started marking derived files to keep 14:32:41.937 -- Started graph log scan Configuration file: /tmp_mnt/shellgame/r/dlusers9/heydon/vesta.cfg Hostname:port: rafael.pa.dec.com:21763 Graph log: /rafael/vesta-srv/cache/graphLog 14:32:43.402 -- Finished graph log scan Graph log entries processed = 1114 Derived files found = 2293 14:32:49.054 -- Finished marking derived files to keep 14:32:55.769 -- Started marking sources and deleting unreachable files 14:44:03.693 -- Finished marking sources and deleting unreachable files
Like most Vesta-2 applications, the cache server reads site-specific configuration information from a Vesta-2 configuration file named vesta.cfg. QuickWeed first looks for this file in the current directory; if none is found there, it looks in your home directory. The -i option disables consultation of the default configuration file, and the -cf option allows you to specify other configuration files.
The configuration file is divided into a number of sections, denoted in the file by [SectionName]. QuickWeed determines which derived files to keep by reading the cache server's graph log. The following variables in the [CacheServer] section of the configuration file are read to determine the location of this file:
- MetaDataRoot (string)
- The pathname of the directory in which the Vesta system's metadata is stored. If this variable is undefined, the current directory is used. Other configuration variables are interpreted relative to this path.
- MetaDataDir (string)
- The directory (relative to the MetaDataRoot) in which the cache server's metadata is stored.
- GraphLogDir (string)
- The directory (relative to the MetaDataRoot/MetaDataDir) in which the cache server stores its graph log.
There are also some variables used from the [Weeder] section of the configuration file:
- MetaDataDir (string)
- The directory (relative to [CacheServer]MetaDataRoot) in which the weeder's metadata is stored.
- Weeded (string)
- The file (relative to [CacheServer]MetaDataRoot/[Weeder]MetaDataDir) in which the set of cache entries to be deleted by the normal weeder (VestaWeed) is stored. QuickWeed checks this as a precaution to make sure that the normal weeder isn't in the middle of doing its work.
- GracePeriod (integer)
- This optional setting specifies an amount of time (in seconds) to subtract from the weeder start time. Any files in the repository (source or derived) changed after that time will be kept during weeding. This closes a small hole which can allow files written by long-running tools that finish close to when the weeder starts to be deleted incorrectly. An ideal setting would be the duration of the longest running tool. Defaults to 60 seconds if not set. Ignored if set to a negative value.
There are no known bugs.
This page was generated automatically by mtex software.
Allan Heydon (email@example.com)