Skip to content

Instantly share code, notes, and snippets.

@Lixivial
Created November 5, 2010 14:43
Show Gist options
  • Star 1 You must be signed in to star a gist
  • Fork 1 You must be signed in to fork a gist
  • Save Lixivial/664236 to your computer and use it in GitHub Desktop.
Save Lixivial/664236 to your computer and use it in GitHub Desktop.
A script to synch files across a set of servers (requires patched wget)
#!/bin/bash
# map_sync.sh - Jesse M. Pearson
# contact: jesall@gmail.com
# or
# irc.quakenet.org/#prae.nexuiz
#
# Description:
# A script to synchronise the latest files across a set of servers (server_list.txt).
# Its intended usage is through a cron task, though manually running it
# one-time via CLI would work, too.
#
# Things to note:
# 1. The script execs wget to do the syncing, however it's been noted that
# if wget -N determines a file is older, it'll check the filesize and
# if that's different, then pull down the file regardless.
#
# Bundled with this script are a set of patchfiles, and compiled binaries
# to disable this feature. If, for some reason, due to architecture or corruption
# or whatever, it should fallback to the system's version of wget.
#
# 2. The files generated/used are mostly self-contained to the script's execution directory.
#
# 3. The default location for parsing relevant information is /tmp/, though $exec_dir would work too
# The point here is that whichever dir is chosen, it needs to be writeable.
#
# TODO:
# 1. Split out the rate limit to become dynamic:
# This can be accomplished by using cURL to get its content-length; specify a max threshold (something by which the rate limit will never exceed)
# and a min threshold (something by which the rate limit will never go below), and dynamically calculate the rate limit to give larger files
# a higher rate limit in order to download them faster.
# 2. Add OS version, and CPU bit support checking to load a proper version wget.
#
# 3. Add an override exclusion routine for stupidly named maps that we want to handle ourselves without the
# ridiculously borked naming conventions.
#
# 4. Clean up the awk/sed's; add a check for $error != "true", so that we can actually
# forego any awk/sedding if we use our custom wget. The custom wget has had its output
# modified to accommodate our file listing.
#
# 5. Add some error handling routines for various commands that might not exist on certain
# configurations.
#
# 6. Var up the extension matches, and use them where ever .pk3 exists. Clean up the script, comment it.
#
# 7. *MAYBE* add some parameter handling so as to control execution flow from cli or cron task
# (i.e. never use custom wget; view verbose wget output in emails; retain wget log; debugging)
# Try to reliably determine the current execution path.
exec_dir=`readlink -f $0`
file_len=`basename $0`
exec_dir=${exec_dir:0:${#exec_dir}-${#file_len}}
date=`date "+%m%d%y%H%M%S"`
server_file="$exec_dir/server_list.txt"
wget="$exec_dir/wget"
parse_file="/tmp/mapsync_$date.txt"
servers=( $( cat $server_file ) )
array_size=$(( ${#servers[@]} ))
directory=$HOME/maps/
extra_params="-T 30 --limit-rate=50k"
# A flag to be set if our custom build of wget borks up due to architecture problems
error="false"
cd ${directory}
for (( i=0; i<array_size; i++ ))
do
# If something has gone wrong with our custom wget build, fallback to system.
if [ $error != "true" ]; then
eval "$wget $extra_params -r -N -nd -A '*.pk3' ${servers[$i]} >> $parse_file 2>&1"
# If something goes wrong with our custom wget build, set the error flag and fallback to system.
if [ $? -gt 0 ]; then
error="true"
echo "Custom wget has failed."
wget $extra_params -r -N -nd -A '*.pk3' ${servers[$i]} >> $parse_file 2>&1
fi
else
wget $extra_params -r -N -nd -A '*.pk3' ${servers[$i]} >> $parse_file 2>&1
fi
done
# This is based on the two different renditions of wget output produced by debian and Mac OS X.
# Hopefully it's more reliable than a static field number. If not, we can cut on .pk3 and rebuild the
# the map name.
maps=$(awk /^[0-9].*\.pk3.*saved/ $parse_file | cut -f2 -d "\`" | cut -f1 -d " " | sed "s/'//")
# This is a dirty hack, I need to figure out how to get awk to give me a more reliable output.
map_count=($(awk /^[0-9].*\.pk3.*saved/ $parse_file | cut -f2 -d "\`" | cut -f1 -d " " | sed "s/'//"))
map_count=${#map_count[@]}
rm $parse_file
if [ $map_count -gt 0 -a "${maps[0]}" != "" ]; then
if [ $map_count -eq 1 ]; then
echo "1 file has been downloaded/updated:"
else
echo "$map_count files have been downloaded/updated:"
fi
echo "${maps}"
#else
# echo "No maps were downloaded/updated."
fi
http://rm.endoftheinternet.org/~nexuiz/maps/
http://maps.nexuizninjaz.com/
http://ouns.nexuizninjaz.com/misc/nexball/maptemp/
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fcpma%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fctf%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fdm%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fmisc%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fnexball%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fnexrun%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fons%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fpb%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Frace%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fcts%2F
http://www.hrfdesign.com/crap/index.php?dir=maps%2Fca%2F
http://www.nullgaming.com/maps/
--- wget-1.11.2/src/http.c 2008-04-27 14:12:26.000000000 -0500
+++ wget-1/src/http.c 2009-09-08 08:43:13.000000000 -0500
@@ -2628,21 +2628,14 @@
download procedure is resumed. */
if (hstat.orig_file_tstamp >= tmr)
{
- if (hstat.contlen == -1
- || hstat.orig_file_size == hstat.contlen)
- {
+ logprintf (LOG_VERBOSE, _("\
+original time stamp `%i' remote time stamp `%i' contlen `%i'.\n\n"),
+ (int)hstat.orig_file_tstamp, (int)tmr, (int)hstat.contlen);
logprintf (LOG_VERBOSE, _("\
Server file no newer than local file `%s' -- not retrieving.\n\n"),
hstat.orig_file_name);
ret = RETROK;
goto exit;
- }
- else
- {
- logprintf (LOG_VERBOSE, _("\
-The sizes do not match (local %s) -- retrieving.\n"),
- number_to_static_string (hstat.orig_file_size));
- }
- }
else
logputs (LOG_VERBOSE,
--- wget-1.11.4/src/http.c 2008-06-29 20:22:54.000000000 -0500
+++ wget-1/src/http.c 2009-09-09 01:56:26.000000000 -0500
@@ -2630,21 +2630,15 @@
download procedure is resumed. */
if (hstat.orig_file_tstamp >= tmr)
{
- if (hstat.contlen == -1
- || hstat.orig_file_size == hstat.contlen)
- {
+ logprintf (LOG_VERBOSE, _("\
+original time stamp `%i' remote time stamp `%i' contlen `%i'.\n\n"),
+ (int)hstat.orig_file_tstamp, (int)tmr, (int)hstat.contlen);
+
logprintf (LOG_VERBOSE, _("\
Server file no newer than local file `%s' -- not retrieving.\n\n"),
hstat.orig_file_name);
ret = RETROK;
goto exit;
- }
- else
- {
- logprintf (LOG_VERBOSE, _("\
-The sizes do not match (local %s) -- retrieving.\n"),
- number_to_static_string (hstat.orig_file_size));
- }
}
else
logputs (LOG_VERBOSE,
--- wget-1.12/src/http.c 2009-09-21 22:02:18.000000000 -0500
+++ wget-1/src/http.c 2009-11-14 09:06:02.000000000 -0600
@@ -2722,21 +2722,11 @@
download procedure is resumed. */
if (hstat.orig_file_tstamp >= tmr)
{
- if (hstat.contlen == -1
- || hstat.orig_file_size == hstat.contlen)
- {
logprintf (LOG_VERBOSE, _("\
Server file no newer than local file %s -- not retrieving.\n\n"),
quote (hstat.orig_file_name));
ret = RETROK;
goto exit;
- }
- else
- {
- logprintf (LOG_VERBOSE, _("\
-The sizes do not match (local %s) -- retrieving.\n"),
- number_to_static_string (hstat.orig_file_size));
- }
}
else
logputs (LOG_VERBOSE,
--- wget-1.12/src/retr.c 2009-09-04 11:31:54.000000000 -0500
+++ wget-1/src/retr.c 2009-11-14 09:06:02.000000000 -0600
@@ -871,8 +871,8 @@
RESTORE_POST_DATA;
bail:
- if (register_status)
- inform_exit_status (result);
+// if (register_status)
+// inform_exit_status (result);
return result;
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment