X-Git-Url: http://git.rot13.org/?p=BackupPC.git;a=blobdiff_plain;f=README.ASA;h=a1f28a2d6a6da1d7b965d40fd885df16a2b665c0;hp=b49aa2adcf8ba6144b4e283a3cc3c0f8561afca4;hb=9fecbd8a0fc6460ca8837c6b38dc9c03ed21c05e;hpb=09666eaeca95859a28b563f36d0babcb3a4390ae

diff --git a/README.ASA b/README.ASA
index b49aa2a..a1f28a2 100644
--- a/README.ASA
+++ b/README.ASA
@@ -2,11 +2,53 @@ This document tries to describe ASA extensions for BackupPC 3.2.0
 
 Written by Dobrica Pavlinusic <dpavlin@rot13.org> 2011-01-27
 
-Search and archive maintain data in PostgreSQL and full-text search. Since full-text search
-is single-writer, we need to serialize somehow requests for it's update.
+This is second iteration of adding search over arbitrary filename substrings and archive
+to CD/DVD media with tracking of copies and additional md5sum creation on them for easy
+burned media verification.
+
+ASA maintains it's data in PostgreSQL and KinoSearch (for faster part-of-filename matching).
+Since full-text search is single-writer, we need to serialize somehow requests for it's update.
+
+Implementation is based on archive host feature in BackupPC using _search_archive.pl configuration
+file located at /etc/BackupPC/pc/_search_archive.pl
+
+This provides us with serialization and hooks around it, but lacked incremental tar creation which
+is essential because we want to burn always increasing archive on CD/DVD media.
+
+This is implemented using new global configuration directive TarCreateIncremental
+
+Using BackupPC hooks to integrate and archive host also provided following advantages:
+	- web interface for archive host contains our log messages
+	- all updates are invoked automatically on end of each run (system is always up to date)
+
+BackupPC can dump multiple machines in parallel, this invoking our _search_archive host and index
+update while update from different machine is still in process. Archive host will reject request,
+but next invocation of same host will fix problem automatically.
+
+To be sure that all pending archives are indexed, you can also run cron job which invokes _search_archive
+on all pending increments:
+
+	/BackupPC_ASA_ArchiveStart _search_archive backuppc
+
+You can also force archival of particual pending backups from single host by adding hostname(s) or
+hostname:num to backup individual increment.
+
+Alternativly, you can use _search_archive web interface to invoke increment creation and indexing.
+
+
+
+There are two global options which had to be set for all hosts:
+
+#
+# /etc/BackupPC/config.pl
+#
+
+# invoke archive of dump - ASA extension DumpPostCmd was too early
+$Conf{DumpPostFinishCmd} = '/srv/BackupPC/bin/BackupPC_ASA_ArchiveStart _search_archive backuppc $host';
+
+# dump only incremental changes in tars not whole content - ASA extension
+$Conf{TarCreateIncremental} = 1;
 
-This is implemented using archive host feature using _search_archive.pl configuration
-file in /etc/BackupPC/pc/_search_archive.pl
 
 
 You can manually trigger all pending backups using:
@@ -15,7 +57,6 @@ You can manually trigger all pending backups using:
 
 This will start archive host _search_archive which will run it's configuration:
 
-
 #
 # /etc/BackupPC/pc/_search_archive.pl
 #
@@ -30,9 +71,6 @@ $Conf{ArchiveDest} = '/data/BackupPC/_search_archive';
 $Conf{ArchiveComp} = 'gzip';
 $Conf{CompressLevel} = 9;
 
-# dump only incremental changes in tars not whole content - ASA extension
-# XXX this option must be global in /etc/BackupPC/config.pl
-$Conf{TarCreateIncremental} = 1;
 
 # archive media size (in bytes) 4.2Gb for DVD
 #$Conf{ArchiveMediaSize} = 4200 * 1024 * 1024; # DVD
@@ -45,14 +83,15 @@ $Conf{ArchiveMediaSize} =  630 * 1024 * 1024; # CD
 # This is useful where the file size of the archive might exceed the
 # capacity of the removable media. For example specify 700 if you are using CDs.
 #$Conf{ArchiveSplit} = 650;
-$Conf{ArchiveSplit} = 42; # FIXME small testing chunks
+$Conf{ArchiveSplit} = 100; # FIXME small testing chunks
 
 
 # The amount of parity data to create for the archive using the par2 utility.
 # In some cases, corrupted archives can be recovered from parity data.
-$Conf{ArchivePar} = 0;
-$Conf{ParPath} = undef;
-
+$Conf{ArchivePar} = 30;
+$Conf{ParPath} = '/srv/par2cmdline-0.4-tbb-20100203-lin64/par2';
+# http://chuchusoft.com/par2_tbb/download.html
+# par2cmdline 0.4 with Intel Threading Building Blocks 2.2
 
 # use parallel gzip (speedup on multi-code machines)
 $Conf{GzipPath} = '/usr/bin/pigz';