fmirror - Mirror directories from ftp servers

fmirror [-a account] [-A and-mask] [-O or-mask] [-C config-line] [-N] [-c dircommand] [-D timezone-dircmd] [-d log-level] [-e exclude-entry] [-i include-entry] [-f config-file] [-F pidfile] [-h] [-k] [-l local dir] [-M use-MDTM] [-m dirmode] [-P port] [-p password] [-r remote dir] [-R] [-s hostname] [-S] [-T remote-timezone] [-t timefuzz] [-u username] [-v] [-V verbosity] [-X maxdel] [-x decompressor] [-o option] [-z]

fmirror is a program for mirroring a directory from a remote ftp server. It allows regex-matching for files that are to be included and excluded. It uses a combination of timestamp, file size and file permissions to decide what files to transfer from the ftp server.

The primary goal of fmirror is to use as little memory as possible, but still be able to do its job efficiently. fmirror can delete any files not found on the remote server.

-a account

Account (not login) id and password separated by forward slash - id/pass. This informations is provided according RFC959.

The argument field is a Telnet string identifying the user's account. The command is not necessarily related to the USER command, as some sites may require an account for login and others only for specific access, such as storing files. In the latter case the command may arrive at any time. There are reply codes to differentiate these cases for the automation: when account information is required for login, the response to a successful PASSword command is reply code 332. On the other hand, if account information is NOT required for login, the reply to a successful PASSword command is 230; and if the account information is needed for a command issued later in the dialogue, the server should return a 332 or 532 reply depending on whether it stores (pending receipt of the ACCounT command) or discards the command, respectively.

-A mask

Remote file permissions will be binary anded by mask before being compared. The same mask is also used when creating files. Default value is 0111, which results in only executable bits being mirrored. The or mask will be ored in after the and mask is anded in.

-O mask

Remote file permissions will be binary ored by mask before being compared. The same mask is also used when creating files. Default value is 0444, which results in all files being readable for everyone. This mask will be ored in after the and mask has been anded in.

-C line

Parses line as if it was a line in a config-file.

-N

Don't delete files that are missing from the ftp server. Files that are replaced by files of other types will still be deleted and updated.

-c command

Use command as a dir-command. To read a ls-lR file instead of performing a heavy ls-operation on the ftp-server, use "RETR ls-lR". The default value is "LIST -lRa".

-D command

Use command as a dir-command for finding the remote timezone. See also the equivalent dircmd_tz: configuration file keyword.

-d debug-level

The level of debugging (spam) output that is wanted. The default debug-level is 3.

-e exclude-opt exclude-pattern

Add an exclude-entry. See discussion for exclude: in the configuration files section. The option and pattern must be one word to the shell, so enclose both in a single "-pair, ie <fmirror -e "p old/">. #"

-i include-opt include-pattern

Add an include-entry. See discussion for include: in the configuration files section.

-f config-file

Reads options from config-file. See the section CONFIGURATION FILES for more information about the contents of a config-file.

-F pid-file

Write the pid (process id) of the fmirror process to pid-file. This is useful when you want to avoid running more than one copy of fmirror from cron or similar.

-h

Display brief option-summary.

-k

Only download files that are newer than the local copy.

-l local-dir

Change directory to local-dir (locally) before mirroring.

-M use-MDTM

set the flag for using MDTM exension. Equivalent to the use_mdtm: keyword, which is described in detail in the configuration file section.

mode. Default value is 0755 (octal).

-m mode

All local directories will be made with mode mode. Default value is 0755 (octal).

-P port

Connect to port port on the remote ftp server.

-p password

The password you want to use on the ftp server. When doing anonymous ftp, this should be your email address. If password is not specified, it be set to 'username@', where username is the username corresponding to the real UID of fmirror.

-r remote-dir

Change directory to remote-dir (on the ftp server) before mirroring.

-R

Set timestamp of local files to the same as remote files (for resyncing the timestamps if you are sure the files are correct). Only sets the timestamps if file lengths and modes match according to the normal rules.

-s hostname

Use hostname as the ftp server to connect to.

-S

Use passive mode for the ftp transfers. This is often necessary to get through some firewalls.

-T remote-timezone

The remote-timezone value (in minutes) overrides the guess fmirror will do on the timezone the remote FTP server operates in. Equivalent to the remotetz: keyword, which is described in detail in the configuration file section.

-t timefuzz

Allow timestamps to differ by up to timefuzz seconds. The default value is 59. No matter what this value is set to, files that differ by exactly 3600 seconds will always considered equal to compensate for buggy timezone-handling. Also, files older than 60 days can differ by up to a day to compensate for lack of precision in ls listings for previous years.

-u username

Sets your username on the ftp server to username. For anonymous ftp, this should be ftp or anonymous. The default value is anonymous.

-v

Show version.

-V verbosity

Verbosity level, 1 is suitable for cron jobs and 3 is suitable for interactive use. Default is 3.

-X maxdel

Maxdel specifies the maximum number of files to delete during fmirror invocation. If the number of deleted files exceeds maxdel the program stops immediately with an error code.

-x program

Run program as a decompressor for the ls listing. This will only be used if compressed dirlisting is on. This should ONLY be the name of the executable, it must not contain any options. The default value is "gzip".

-o option

Give option as option to the decompressor. The default value is "-dc".

-z

Enable decompression of dir-listing.

Unless you are just using this program to mirror a directory once, you'll probably want to make a configuration file to make the job easier.

The syntax of the configuration file is very simple. Any line not starting with an alphabetical letter is considered a comment line (blank lines are also considered comments).

The first option of a particular kind has the highest priority, and configuration file and command line options have the same priority. You can have any number of include and exclude options, the first one that matches a particular file will be used.

You can also include multiple configuration files, the first configuration file has precedence over later ones. The debug-level can be changed between every single option on the command-line to debug specific configuration files.

NOTE: All numerical values can be written in octal (leading 0), decimal (default) or hexadecimal (leading 0x). Modes are usually written in octal, so make sure you write 0755 and not 755 (for example).

username: username

Sets the username (same as -u command line option). Any whitespace before the username will be ignored. This will typically be "ftp" or "anonymous".

password: password

Sets the password (same as -p command line option). Any whitespace before the password will be ignored. This will typically be "user@some.domain", ie "finnag@fast.no"

host: remote hostname

Specifies the remote hostname (same as -s command line option). Any whitespace before the hostname will be ignored.

remotedir: remote directory

Specifies the remote directory that is to be mirrored. (same as -r command line option). Any whitespace before the name will be ignored.

localdir: local directory

Specifies the local directory that you are mirroring to (same as -l command line option). Any whitespace before the name will be ignored.

dircmd: directory list command

Specifies what command should be sent to the ftp-server to get a directory listing. Normally this is "LIST -lRa". If you want to get a ls-lR file instead, you can put "RETR ls-lR" here. This option is equivalent to the -c command line option.

dircmd_tz: directory list command

Specifies what command should be sent to the ftp-server to get a directory listing when trying to find the remote timezone. The default value is "LIST -ltra". Should not be set to any recursive command. Equivalent to the -D command line option.

decompressor:

If you are getting a compressed file-listing, this should be the name of the decompressor program without any options. This is set to "gzip" if you only enable compressed dir-listings without specifying a decompressor. Equivalent to -x command line option.

decompressor_opt:

The option string that is to be sent to the decompressor. This defaults to "-dc". Equivalent to -o command line option.

port: port

Connect to port port on the remote ftp server. Equivalent to -P command line option.

pidfile: pid-file

Store the PID (process id) of the process to the file pid-file. Equivalent to -F command line option.

dirmode: mode

All local directories will be created with mode mode. Equivalent to -m command line option.

loglevel: level

Specifies the logging level (debug level). Equivalent to -d command line option.

compressed: val

Takes an integer argument, will enable decompressing of the file listing if val is non-zero. With a non-zero value this is equivalent to the -z command line option.

timefuzz: fuzz

The maximum number of seconds files can differ by and still be considered equal. Equivalent to -t command line option.

file_and_mask: mask

File permissions will be anded by mask before being compared and created. Equivalent to -A command line option.

file_or_mask: mask

File permissions will be ored by file_or_mask before being created. Equivalent to -O command line option.

passive: val

Use passive mode if val is non-zero, or normal mode if it is zero.

nodel: val

Don't delete local files if val is non-zero (equivalent to -N command line option). Normal operation if val is zero.

log_timestamp: val

Only prepend timestamp on each line output if val is non-zero. Default value is 1.

verbosity: verbosity

Chose verbosity value. The only supported values are 1, which is suitable for non-interactive operation, and 3, which is suitable for interactive operation.

keep_newer: val

Takes an integer argument, will keep newer file if val is non-zero. This means that remote files that have older timestamps than local files will not cause that file to be retrieved. Equivalent to -k command line option.

reset_times: val

Takes an integer argument, will reset times of files if val is non-zero. This means that files thay only differ in time-stamp, but match in size and mode, will get their local timestamp adjusted to that of the remote files instead of being retrieved again. Equivalent to -R command line option.

remotetz: val

Takes an integer argument. The val is the remote timezone in minutes. Typical values would be -60 for most of Europe, -540 for Japan, anywhere from 300 to 480 for the continental US, and 600 for Hawaii. Normally, fmirror will only try to guess the timezone when use_mdtm is in effect, and assume GMT in all other cases. Overriding the time zone guess is particularly important when you are fetching a ls-lR file with RETR instead of actually doing a LIST, since the guess is based on doing LIST and MDTM and comparing the results.

use_mdtm: val

Takes an integer argument. The val decides whether to try using the MDTM extension for getting remote modification times. To avoid constantly regetting files, this also needs to figure out what timezone the remote server is running in, by doing a remote "LIST -ltra", picking a file with time information, sending an MDTM command for it, and comparing the results. See also the dircmd_tz: and remotetz: keywords.

A value of 0 turns off trying MDTM, a value of 1 means to use it if it is available and we can determine the remote timezone, while any other value (preferably 2) means to use it if it is available and assume that the remote timezone is GMT if we can't determine it normally. In the last case, using a large value for timefuzz (like 50000 for a bit more than 12 hours) may also be useful. The default value is 1.

input_timeout: num

Will wait num seconds before forcing a read of a file to time out. No equivalent command line option, default value is 240 seconds.

connect_timeout: num

The timeout before retrying a connect operation (not the initial connect, but the connects that are required to transfer files after the control connection is up). Will wait num seconds before timing out on connects (or accepts for non-passive ftp transfers). No equivalent command line option, default value is 120 seconds.

connect_retries: num

The number of times to retry a connect operation (not the initial connect, but the connects that are required to transfer files after the control connection is up). Will try to connect up to num times before giving up. No equivalent command line option, default value is 3.

reconnect_timeout: num

Will wait num seconds before retrying a connect operation to a ftp server, either if there was no reply from the ftp server at all at the initial connect, or if the username/password combo was not allowed (usually happens on full ftp servers). No equivalent command line option, default value is 120 seconds.

reconnect_retries: num

Will try to connect up to num times before giving up if initial connection cannot be made. This means that eitehr the ftp servre does not reply at all, or fmirror is kicked out because the ftp server is full. No equivalent command line option, default value is 15;

maxdel: num

Maxdel specifies the maximum number of files to delete during fmirror invocation. If the number of deleted files exceeds num the program stops immediately with an error code.

exclude: [i][x][n]{p|f} pattern

A regex pattern that, if matched in a particular file/path-name, should not be mirrored. This will override any include-patterns encountered after this. Equivalent to -e command line option.

NOTE: An exclude-rule doesn't differ between files or directories. p or f only decide whether the entire path-name or just the name of the file/directory itself should be matched. Directories get "/" appended before they are matched against the regex, so "exclude f ^old/" will exclude a directory named "old" (but not necessarily its contents), but not a file named "old".

Directories are created "on demand", so if you have excluded "old/" but not "\.gif$", a file with path "old/bill.gif" will cause the "old"-directory to be made and the "bill.gif"-file to be retrieved. To exclude any directory named "old" and all its contents, use "p (^|/)old/".

i means case insensitive match.

x means to match anything EXCEPT the following regex, effectively inverting the rule.

n means that the file/directory will not be deleted if it exists locally, matches this rule, and would otherwise be deleted.

p means match full path-name (eg "pub/foo.bar").

f means match filename only (eg "foo.bar")

pattern is an extended POSIX regular expression.

include: [i][x][n]{p|f} pattern

A regex pattern that, if matched in a particular file/path-name, should be mirrored. This will override any exclude-patterns encountered after this. Equivalent to -i command line option.

The options are identical with the exclude options.

There should be several sample configuration files with your copy of fmirror in the configs directory.

Transient errors are not handled well. If one file cannot be transferred for some reason, such as 'Failed to make data socket', it doesn't try to get that file again, so you effectively lose the file.

Copyright 1995-2000 Finn Arne Gangstad and Tor Egge.

Permission is granted to make and distribute verbatim copies of this manual provided the copyright notice and this permission notice are preserved on all copies.

Permission is granted to copy and distribute modified versions of this manual under the conditions for verbatim copying, provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one.

Permission is granted to copy and distribute translations of this manual into another language, under the above conditions for modified versions, except that this permission notice may be included in translations approved by the Free Software Foundation instead of in the original English.

fmirror was written by

Finn Arne Gangstad <finnag@fast.no>