From 544a7d269ec9d4546b03879b18bab7180636a3e8 Mon Sep 17 00:00:00 2001
From: jll2 <jll2_8854a@redwoodscientific.com>
Date: Wed, 24 Jun 2009 17:01:14 -0700
Subject: [PATCH] Create rsync_failure_codes option.

User may optionally create a file rsync_failure_codes with a
newline-separated list of rsync return codes that are to be
regarded as complete failure.

If rsync exits with such a code, then the backup will be marked
for deletion during the next ccollect run (if delete_incomplete).

I added documentation for this feature in doc/ccollect.text

In my experience (yours may differ), two rsync exit codes that
belong in this file are 12 and 255.  I have seen 12 result from
ssh errors and 255 result from a kernel module conflict.  In both
cases, the resulting backups were empty.  Without the
rsync_failure_codes option, such errors cause good backups to be
deleted, as per c_interval, leaving behind the new empty backups.

In the long run, we may want a different and more comprehensive
method for analyzing rsync errors.  In the short run, I find this
option necessary.
---
 ccollect.sh       | 36 ++++++++++++++++++++++++++----------
 doc/ccollect.text | 10 ++++++++++
 2 files changed, 36 insertions(+), 10 deletions(-)

diff --git a/ccollect.sh b/ccollect.sh
index d6831bc..2b67ba7 100755
--- a/ccollect.sh
+++ b/ccollect.sh
@@ -281,7 +281,8 @@ while [ "${i}" -lt "${no_sources}" ]; do
    c_dest="${backup}/destination"
    c_pre_exec="${backup}/pre_exec"
    c_post_exec="${backup}/post_exec"
-   for opt in exclude verbose very_verbose rsync_options summary delete_incomplete remote_host ; do
+   for opt in exclude verbose very_verbose rsync_options summary delete_incomplete \
+         remote_host rsync_failure_codes ; do
       if [ -f "${backup}/$opt" -o -f "${backup}/no_$opt"  ]; then
          eval c_$opt=\"${backup}/$opt\"
       else
@@ -530,16 +531,31 @@ while [ "${i}" -lt "${no_sources}" ]; do
 
    _techo "Transferring files..."
    rsync "$@" "${source}" "${destination_full}"; ret=$?
-
-   #
-   # remove marking here
-   #
-   pcmd rm "${destination_dir}.${c_marker}" || \
-      _exit_err "Removing ${destination_dir}/${c_marker} failed."
-
    _techo "Finished backup (rsync return code: $ret)."
-   if [ "${ret}" -ne 0 ]; then
-      _techo "Warning: rsync exited non-zero, the backup may be broken (see rsync errors)."
+
+   #
+   # Check if rsync exit code indicates failure.
+   #
+   fail=""
+   if [ -f "$c_rsync_failure_codes" ]; then
+      while read code ; do
+         if [ "$ret" = "$code" ]; then
+            fail=1
+         fi
+      done <"$c_rsync_failure_codes"
+   fi
+
+   #
+   # Remove marking here unless rsync failed.
+   #
+   if [ -z "$fail" ]; then
+      pcmd rm "${destination_dir}.${c_marker}" || \
+         _exit_err "Removing ${destination_dir}/${c_marker} failed."
+      if [ "${ret}" -ne 0 ]; then
+         _techo "Warning: rsync exited non-zero, the backup may be broken (see rsync errors)."
+      fi
+   else
+      _techo "Warning: rsync failed with return code $ret."
    fi
 
    #
diff --git a/doc/ccollect.text b/doc/ccollect.text
index f1451f8..3d9c480 100644
--- a/doc/ccollect.text
+++ b/doc/ccollect.text
@@ -359,6 +359,7 @@ Additionally a source may have the following files:
    
    - `delete_incomplete`   delete incomplete backups
    - `remote_host`         host to backup to
+   - `rsync_failure_codes` list of rsync exit codes that indicate complete failure
 
 
 Example:
@@ -574,6 +575,15 @@ If you create the file `delete_incomplete` in a source specification directory,
 was interrupted) and remove them. Without this file `ccollect` will only warn
 the user.
 
+Detailed description of "rsync_failure_codes"
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+If you have the file `rsync_failure_codes` in your source configuration
+directory, it should contain a newline-separated list of numbers representing
+rsync exit codes.  If rsync exits with any code in this list, a marker will
+be left in the destination directory indicating failure of this backup.  If
+you have enabled delete_incomplete, then this backup will be deleted during
+the next ccollect run on the same interval.
+
 
 Hints
 -----