From 94c5c402c64347efa26a6d358cdc7ad4bb0146ef Mon Sep 17 00:00:00 2001
From: "Kyle J. McKay" <mackyle@gmail.com>
Date: Fri, 28 Nov 2014 15:42:27 -0800
Subject: [PATCH] reports: new project-disk-use.sh report

Reports total disk used by projects sorted in descending piggish order.
---
 cron/girocco                        |  3 ++
 install.sh                          |  4 +-
 toolbox/{inactive => inactive.sh}   |  0
 toolbox/reports/project-disk-use.sh | 96 +++++++++++++++++++++++++++++++++++++
 4 files changed, 102 insertions(+), 1 deletion(-)
 rename toolbox/{inactive => inactive.sh} (100%)
 create mode 100755 toolbox/reports/project-disk-use.sh

diff --git a/cron/girocco b/cron/girocco
index 4d786e8..97c94cb 100644
--- a/cron/girocco
+++ b/cron/girocco
@@ -53,6 +53,9 @@
 # Nine rotating backups are kept, run this more often if needed.
 19 5 * * *	repo	"$HOME/repomgr/toolbox/backup-db.sh"
 
+# Report project disk space usage once a week with an email to admin
+#23 7 * * 3	repo	"$HOME/repomgr/toolbox/reports/project-disk-use.sh" -m 100
+
 # The job daemon and task daemon are run in a screen
 # The ../screen/screenrc file needs to be installed to
 # ~repo/.screenrc to support the SCREENCOMMAND functionality.
diff --git a/install.sh b/install.sh
index f4581b0..880abc7 100755
--- a/install.sh
+++ b/install.sh
@@ -125,7 +125,9 @@ perl -I. -M$GIROCCO_CONF -i -p \
 	"$cfg_basedir"/jobs/*.sh "$cfg_basedir"/jobd/*.sh \
 	"$cfg_basedir"/taskd/*.sh "$cfg_basedir"/gitweb/*.sh \
 	"$cfg_basedir"/shlib.sh "$cfg_basedir"/hooks/* \
-	"$cfg_basedir"/toolbox/* "$cfg_basedir"/bin/git-* \
+	"$cfg_basedir"/toolbox/*.sh "$cfg_basedir"/toolbox/*.pl \
+	"$cfg_basedir"/toolbox/reports/*.sh \
+	"$cfg_basedir"/bin/git-* \
 	"$cfg_basedir"/bin/create-* "$cfg_basedir"/bin/update-* \
 	"$cfg_basedir"/bin/authrequired.cgi "$cfg_basedir"/screen/*
 
diff --git a/toolbox/inactive b/toolbox/inactive.sh
similarity index 100%
rename from toolbox/inactive
rename to toolbox/inactive.sh
diff --git a/toolbox/reports/project-disk-use.sh b/toolbox/reports/project-disk-use.sh
new file mode 100755
index 0000000..2a761fa
--- /dev/null
+++ b/toolbox/reports/project-disk-use.sh
@@ -0,0 +1,96 @@
+#!/bin/sh
+
+# Report on project disk use.
+# Output can be sent as email to admin with -m
+# Automatically runs with nice and ionice (if available)
+
+# Usage: project-disk-use [-m] [top-n-only]
+
+# With -m mail the report to $cfg_admin instead of sending it to stdout
+
+# Shows total disk usage in K bytes for $cfg_reporoot
+# The top-n-only (all if not given) repos are then listed by
+# decreasing order of disk space.
+
+# Note that any *.git directories that are found in $cfg_reporoot that are
+# not listed in $cfg_chroot/etc/group ARE included and have a '!' suffix added.
+
+set -e
+
+datefmt='%Y-%m-%d %H:%M:%S %z'
+startdate="$(date "+$datefmt")"
+
+. @basedir@/shlib.sh
+
+mailresult=
+if [ "$1" = "-m" ]; then
+	shift
+	mailresult=1
+fi
+
+hasnice=
+command -v nice > /dev/null && hasnice=1
+hasionice=
+command -v ionice > /dev/null && hasionice=1
+
+get_use_k() {
+	_cmd="du -k -s \"$1\" | cut -f 1"
+	[ -n "$hasionice" ] && _cmd="ionice -c 3 $_cmd"
+	[ -n "$hasnice" ] && _cmd="nice -n 19 $_cmd"
+	eval "$_cmd"
+}
+
+projlist="$(cut -d : -f 1 < "$cfg_chroot/etc/group")"
+
+is_listed_proj() {
+  echo "$projlist" | grep -q -e "^$1$"
+}
+
+cd "$cfg_reporoot"
+total=0
+results=
+while IFS='' read -r proj; do
+	proj="${proj#./}"
+	proj="${proj%.git}"
+	x=
+	is_listed_proj "$proj" || x='!'
+	usek="$(get_use_k "$proj.git")"
+	total="$(( $total + $usek ))"
+	line="$(printf '%8d\t%s%s\n*' "$usek" "$proj" "$x")"
+	line="${line%?}"
+	results="$results$line"
+done <<EOT
+$(find . -type d -name '*.git' -print -prune)
+EOT
+
+enddate="$(date "+$datefmt")"
+domail=cat
+[ -n "$mailresult" ] && domail='mail -s "[$cfg_name] Project Disk Use Report" "$cfg_admin"'
+{
+	cat <<EOT
+Project Disk Use Report
+=======================
+
+Repository Root: $cfg_reporoot
+     Start Time: $startdate
+       End Time: $enddate
+ Total Disk Use: $total (1024-byte blocks)
+
+$(df -h "$cfg_reporoot")
+EOT
+
+	if [ $# -lt 1 ] || [ $1 != "0" ]; then
+		topn=cat
+		message="Individual Repository Use"
+		case "${1%%[!0-9]*}" in ?*)
+			message="Individual Repository Use (top $1)"
+			topn="head -n ${1%%[!0-9]*}"
+		esac
+		echo ""
+		echo ""
+		echo "$message"
+		echo "$message" | tr -c '\n' -
+		echo ""
+		printf '%s' "$results" | sort -k1,1nr -k2,2 | $topn
+	fi
+} | eval "$domail"
-- 
2.11.4.GIT