holm/zfs-auto-snapshot
1
0
Fork 0
forked from extern/zfs-auto-snapshot
Code Issues Pull Requests Packages Projects Releases Wiki Activity

check_failure always says it's moving to maintenance mode (even when it doesn't actually)

Browse Source 
Source code comments cleanup
This commit is contained in:
Tim Foster 2008-10-10 11:51:32 +01:00
parent 644b686ebe
commit c24c6b515c
3 changed files with 32 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

56
src/lib/svc/method/zfs-auto-snapshot
View File

@ -27,15 +27,14 @@
# zfs-auto-snapshot changeset = ~ZFS_AUTO_SNAPSHOT_CHANGESET~ # zfs-auto-snapshot changeset = ~ZFS_AUTO_SNAPSHOT_CHANGESET~
# This SMF method takes periodic snapshots of zfs filesystems, with
# # options to allow the user to keep a limited number of snapshots, snapshot
# This SMF method takes snapshots periodically of a zfs filesystem, with # all child datasets, or run zfs send piping to a specified command.
# options to allow the user to keep a limited number of snapshots, or snapshot # More documentation is available at
# all child datasets. More documentation available at # http://src.opensolaris.org/source/xref/jds/zfs-snapshot/README.zfs-auto-snapshot.txt
# http://blogs.sun.com/timf
# #
# The service will move itself into maintenance if it's unable to take a # The service will move itself into maintenance if it's unable to take a
# snapshot, destroy a snapshot as per the snapshot retention policy, unable to # snapshot, destroy a snapshot as per the snapshot retention policy, is unable to
# zfs send a dataset (if configured) or is unable to create or update the cron # zfs send a dataset (if configured) or is unable to create or update the cron
# job. # job.
# #
@ -49,14 +48,13 @@
# subprocesses, and when called with a non-zero argument, will degrade the # subprocesses, and when called with a non-zero argument, will degrade the
# service, and log an appropriate error message. # service, and log an appropriate error message.
. /lib/svc/share/smf_include.sh . /lib/svc/share/smf_include.sh
result=$SMF_EXIT_OK result=$SMF_EXIT_OK
export PATH=/usr/sbin:/usr/bin:${PATH} export PATH=/usr/sbin:/usr/bin:${PATH}
# A prefix we use on all snapshot created by this script. # A prefix we use on all snapshots created by this script.
# See the definition of $SNAPNAME in the take_snapshot() # See the definition of $SNAPNAME in the take_snapshot()
# function for more information. # function for more information.
PREFIX="zfs-auto-snap" PREFIX="zfs-auto-snap"
@ -71,10 +69,9 @@ SEP=":"
# and use logger(1) to log to syslog instead. # and use logger(1) to log to syslog instead.
LOG="" LOG=""
# Determine whether this invocation of the script can use # Determine whether this invocation of the script can use
# the recursive snapshot feature in some versions of ZFS. # the recursive snapshot feature in some versions of ZFS.
# a null string in this variable says we don't. # A null string in this variable says we don't.
HAS_RECURSIVE=$(zfs snapshot 2>&1 | fgrep -e '-r') HAS_RECURSIVE=$(zfs snapshot 2>&1 | fgrep -e '-r')
# OpenSolaris has no pfksh, so for now, we need to pfexec # OpenSolaris has no pfksh, so for now, we need to pfexec
@ -90,6 +87,7 @@ function get_pair {
# A function to pull in the variables from the FMRI given # A function to pull in the variables from the FMRI given
# as the first argument. # as the first argument.
#
function zfs_smf_props { function zfs_smf_props {
IFS=" IFS="
@ -113,6 +111,7 @@ done
# calls a function to create cron job that schedules a snapshot schedule based # calls a function to create cron job that schedules a snapshot schedule based
# on the properties set in the service instance. # on the properties set in the service instance.
# $1 is assumed to be a valid FMRI # $1 is assumed to be a valid FMRI
#
function schedule_snapshots { function schedule_snapshots {
typeset FMRI=$1 typeset FMRI=$1
@ -167,11 +166,10 @@ function schedule_snapshots {
} }
#
# Adding a cron job that runs exactly every x time-intervals is hard to do # Adding a cron job that runs exactly every x time-intervals is hard to do
# properly. # properly.
# #
# For now, what I'm doing, is dividing the interval up into x bite chunks # For now, what we're doing, is dividing the interval up into x bite chunks
# and running the cron job that often. The problem comes where the interval # and running the cron job that often. The problem comes where the interval
# doesn't evenly divide up into x, leaving us taking to many, or too # doesn't evenly divide up into x, leaving us taking to many, or too
# few snapshots at the edge of time intervals. # few snapshots at the edge of time intervals.
@ -215,7 +213,7 @@ function add_cron_job { # $INTERVAL $PERIOD $OFFSET $FMRI
# Since we may have multiple instances all trying to start at # Since we may have multiple instances all trying to start at
# the same time, we need some form of locking around crontab. # the same time, we need some form of locking around crontab.
# Normally we'd be able to get SMF to manage this, by defining dependencies - # Normally we'd be able to get SMF to manage this, by defining dependencies -
# but I'm not sure there's a way to prevent it from starting two instances # but we're not sure there's a way to prevent it from starting two instances
# at the same time (without requiring users to explicitly state dependencies # at the same time (without requiring users to explicitly state dependencies
# and change them each time new instances are added) # and change them each time new instances are added)
@ -232,7 +230,7 @@ function add_cron_job { # $INTERVAL $PERIOD $OFFSET $FMRI
done done
# adding a cron job is essentially just looking for an existing entry, # adding a cron job is essentially just looking for an existing entry,
# removing it, and appending a new one. Neato. # removing it, and appending a new one.
crontab -l | grep -v "/lib/svc/method/zfs-auto-snapshot $FMRI$" \ crontab -l | grep -v "/lib/svc/method/zfs-auto-snapshot $FMRI$" \
> /tmp/saved-crontab.$$ > /tmp/saved-crontab.$$
@ -252,6 +250,7 @@ function add_cron_job { # $INTERVAL $PERIOD $OFFSET $FMRI
# this function removes a cron job was taking snapshots of $FMRI # this function removes a cron job was taking snapshots of $FMRI
# $1 is assumed to be a valid FMRI # $1 is assumed to be a valid FMRI
#
function unschedule_snapshots { function unschedule_snapshots {
typeset FMRI=$1 typeset FMRI=$1
@ -295,6 +294,7 @@ function unschedule_snapshots {
# This function is intended to be called on service start. It checks to see # This function is intended to be called on service start. It checks to see
# if the last snapshot was taken more than <frequency> <intervals> ago, # if the last snapshot was taken more than <frequency> <intervals> ago,
# and if that's the case, takes a snapshot immediatedly. # and if that's the case, takes a snapshot immediatedly.
#
function check_missed_snapshots { # $INTERVAL $PERIOD $FMRI <repopulate cache> function check_missed_snapshots { # $INTERVAL $PERIOD $FMRI <repopulate cache>
typeset INTERVAL=$1 typeset INTERVAL=$1
@ -308,7 +308,7 @@ function check_missed_snapshots { # $INTERVAL $PERIOD $FMRI <repopulate cache>
fi fi
# // is special, in that we take snapshots based on user properties # // is special, in that we take snapshots based on user properties
# so here, we get those properties, and call ourselves again, with # so here, we get those properties and call ourselves again, with
# those values. # those values.
case "$FILESYS" in case "$FILESYS" in
"//") "//")
@ -345,14 +345,13 @@ function check_missed_snapshots { # $INTERVAL $PERIOD $FMRI <repopulate cache>
LAST_SNAPSHOT=$(zfs list -H -o name -r -t snapshot ${fs[0]} \ LAST_SNAPSHOT=$(zfs list -H -o name -r -t snapshot ${fs[0]} \
| grep "${fs[0]}@${PREFIX}${LABEL}" | tail -1) | grep "${fs[0]}@${PREFIX}${LABEL}" | tail -1)
# if we've never taken a snapshot, then we probably should # if we've never taken a snapshot, then we should
# arrange to have one taken now. Fake the LAST_SNAP_TIME variable # arrange to have one taken now. Fake the LAST_SNAP_TIME variable
if [ -z "$LAST_SNAPSHOT" ] ; then if [ -z "$LAST_SNAPSHOT" ] ; then
LAST_SNAP_TIME=0 LAST_SNAP_TIME=0
LAST_SNAP_TIME_HUMAN="[ no snapshot taken yet ]" LAST_SNAP_TIME_HUMAN="[ no snapshot taken yet ]"
fi fi
# if we have taken a snapshot, find out when it was # if we have taken a snapshot, find out when it was
if [ -n "$LAST_SNAPSHOT" ] ; then if [ -n "$LAST_SNAPSHOT" ] ; then
LAST_SNAP_TIME=$(zfs get -H -p -o value creation $LAST_SNAPSHOT) LAST_SNAP_TIME=$(zfs get -H -p -o value creation $LAST_SNAPSHOT)
@ -419,14 +418,14 @@ function take_snapshot {
# requested one, which then gets used in the SNAPNAME. SMF # requested one, which then gets used in the SNAPNAME. SMF
# returns the value '""' for the empty string to differentiate # returns the value '""' for the empty string to differentiate
# between an unset property, and a set-but-empty property. # between an unset property, and a set-but-empty property.
# Shocking, I know.
typeset LABEL="$label" typeset LABEL="$label"
# the "//" filesystem is special. We use it as a keyword # the "//" filesystem is special. We use it as a keyword
# to determine whether to poll the ZFS "com.sun:auto-snapshot:${LABEL}" # to determine whether to poll the ZFS "com.sun:auto-snapshot:${LABEL}"
# user property which specifies which datasets should be snapshotted # user property which specifies which datasets should be snapshotted
# and under which "label" - a set of default service instances that # and under which "label" - a set of default service instances that
# snapshot at defined periods (daily, weekly, monthly, every 15 mins) # snapshot at defined periods (daily, hourly, weekly, monthly, every
# 15 mins) are provided already.
# Determine what these are, call ourselves again, then return. # Determine what these are, call ourselves again, then return.
case "$FILESYS" in case "$FILESYS" in
"//") "//")
@ -454,19 +453,19 @@ function take_snapshot {
LABEL="" LABEL=""
fi fi
# A flag for whether we're running in verbose mode or not # A flag for whether we're running in verbose mode
VERBOSE="$verbose" VERBOSE="$verbose"
typeset SNAPNAME="${PREFIX}${LABEL}-${DATE}" typeset SNAPNAME="${PREFIX}${LABEL}-${DATE}"
# Determine whether we should avoid scrubbing # Determine whether we should avoid pools which are scrubbing
typeset AVOIDSCRUB=$avoidscrub typeset AVOIDSCRUB=$avoidscrub
# prune out the filesystems that are on pools currently being # prune out the filesystems that are on pools currently being
# scrubbed or resilvered. There's a risk that a scrub/resilver # scrubbed or resilvered. There's a risk that a scrub/resilver
# will be started just after this check completes, but there's # will be started just after this check completes, and also
# also the risk that a running scrub will complete just after this # the risk that a running scrub will complete just after this
# check. Life's hard. # check.
if [ "$AVOIDSCRUB" == "true" ] ; then if [ "$AVOIDSCRUB" == "true" ] ; then
# lists of pools and datasets that are known not to be scrubbing # lists of pools and datasets that are known not to be scrubbing
NOSCRUBLIST="" NOSCRUBLIST=""
@ -578,7 +577,7 @@ function destroy_older_snapshots {
# we don't degrade the service. We also log an error message as passed into # we don't degrade the service. We also log an error message as passed into
# this function. # this function.
# #
function check_failure { # integer exit status, error message to display, be fatal function check_failure { # integer exit status, error message, non-fatal flag
typeset RESULT=$1 typeset RESULT=$1
typeset ERR_MSG=$2 typeset ERR_MSG=$2
@ -586,7 +585,6 @@ function check_failure { # integer exit status, error message to display, be fat
if [ $RESULT -ne 0 ] ; then if [ $RESULT -ne 0 ] ; then
print_log "Error: $ERR_MSG" print_log "Error: $ERR_MSG"
print_log "Moving service $FMRI to maintenance mode."
if [ -z "${NON_FATAL}" ] ; then if [ -z "${NON_FATAL}" ] ; then
print_log "Moving service $FMRI to maintenance mode." print_log "Moving service $FMRI to maintenance mode."
svcadm mark maintenance $FMRI svcadm mark maintenance $FMRI
@ -598,7 +596,8 @@ function check_failure { # integer exit status, error message to display, be fat
# A function we use to emit output. Right now, this goes to syslog via logger(1) # A function we use to emit output. Right now, this goes to syslog via logger(1)
# as well as being echoed to stdout which will result in it being picked up by # as well as being echoed to stdout which will result in it being picked up by
# SMF. # SMF if the $LOG variable is unset.
#
function print_log { # message to display function print_log { # message to display
logger -t zfs-auto-snap -p daemon.notice $* logger -t zfs-auto-snap -p daemon.notice $*
if [ -z "$LOG" ] ; then if [ -z "$LOG" ] ; then
@ -609,6 +608,7 @@ function print_log { # message to display
# Another function to emit output, this time checking to see if the # Another function to emit output, this time checking to see if the
# user has set the service into verbose mode, otherwise, we print nothing # user has set the service into verbose mode, otherwise, we print nothing
# This goes to stdout, and will get collected by either SMF or cron # This goes to stdout, and will get collected by either SMF or cron
#
function print_note { # mesage to display function print_note { # mesage to display
if [ "$VERBOSE" == "true" ] ; then if [ "$VERBOSE" == "true" ] ; then
echo $(date "+%b %d %T") $* echo $(date "+%b %d %T") $*

2
src/pkginfo
View File

@ -3,7 +3,7 @@ LANG=C
TZ=Eire TZ=Eire
PATH=/sbin:/usr/sbin:/usr/bin:/usr/sadm/install/bin PATH=/sbin:/usr/sbin:/usr/bin:/usr/sadm/install/bin
OAMBASE=/usr/sadm/sysadm OAMBASE=/usr/sadm/sysadm
SITENAME=Sun Microsystems SITENAME=Sun Microsystems, Inc.
PKG=SUNWzfs-auto-snapshot PKG=SUNWzfs-auto-snapshot
NAME=ZFS Automatic Snapshot Service NAME=ZFS Automatic Snapshot Service
ARCH=all ARCH=all

8
src/var/svc/manifest/system/filesystem/auto-snapshot.xml
View File

@ -34,8 +34,7 @@
type='service' type='service'
version='0.11'> version='0.11'>
<!-- we need to be multi-user to ensure the homedirs for the <!-- we need to be multi-user -->
zfssnap role is mounted -->
<dependency <dependency
name='multi-user' name='multi-user'
grouping='require_all' grouping='require_all'
@ -94,8 +93,7 @@ they being :
properties on datasets, set to "true" if the dataset properties on datasets, set to "true" if the dataset
should have snapshots taken by this instance. should have snapshots taken by this instance.
If set to false or unset, snapshots will not be taken If unset, snapshots will not be taken by this instance.
by this instance.
The snapshot-children property is ignored when using The snapshot-children property is ignored when using
this setting, instead the system will automatically this setting, instead the system will automatically
@ -109,7 +107,7 @@ they being :
interval : minutes | hours | days | months | none interval : minutes | hours | days | months | none
For the interval "none" a cron job is not created for that For the interval "none" a cron job is not created for that
instance - instead the user can manually file the method instance - instead the user can manually fire the method
script to take snapshots defined by the rest of the properties script to take snapshots defined by the rest of the properties
in the instance. The period and offset values are ignored in in the instance. The period and offset values are ignored in
this case. this case.