GitHub/ansitest
1
0
Fork 0
mirror of https://github.com/kneutron/ansitest.git synced 2025-01-16 04:42:55 +08:00
Code Issues Packages Projects Releases Wiki Activity
ansitest/ZFS/move-home-to-zfs.sh
kneutron 531e0e5926
ZFS utility / admin scripts
2021-04-12 14:40:45 -05:00

358 lines
13 KiB
Bash
Raw Blame History

#!/bin/bash
# TODO - sep datasets for users
# DONE new disk needs to be at least = size of /home du -s -h
# DONE free snapshot
# GOAL: move existing /home to zfs
# $1 = disk name (long or short) OR existing zfs poolname
# =LLC= © (C)opyright 2016 Boojum Consulting LLC / Dave Bechtel, All rights reserved.
## NOTICE: Only Boojum Consulting LLC personnel may use or redistribute this code,
## Unless given explicit permission by the author - see http://www.boojumconsultingsa.com
#
logfile=~/boojum-mvhome2zfs.log
source ~/bin/logecho.mrg
> $logfile
# If set to 1, will interactively kill user processes
# EITHER of these options (if set) will destroy existing ZFS datasets!
debugg=0
RESETALL=0
# DANGEROUS - ONLY SET IF U KNOW WHAT U DOING AFTER RESTORING A SNAPSHOT!
# WILL DESTROY $zp
# TODO edit this to be the name of the ZFS home pool you want to be created, if needed
zp=zhome
[ $RESETALL -gt 0 ] && (set -x;zpool destroy $zp)
tmpfile1=/tmp/mvh2zTMP1.txt
source ~/bin/failexit.mrg
modprobe zfs # shouldnt hurt even if its already loaded
# is zfs installed?
#zfsps=`zpool status |head -n 1`
zfsps=$(zfs list |head -n 1)
if [ `echo "$zfsps" |grep -c 'MOUNTPOINT'` -ge 1 ]; then
logecho 'Existing zfs pool(s) detected:'
zpool status |awk 'NF>0'
echo 'FYI: Pass a ZFS pool name to this script to move /home there, or pass a disk name to create a new pool'
elif [ `echo "$zfsps" |grep -c 'no datasets available'` ]; then
logecho "NOTE: ZFS is installed and appears to be working - will create a pool ( $zp ) to hold /home"
else
logecho '! ZFS does not appear to be installed or is not working correctly'
failexit 99 '! zpool status is not returning a valid result:'
(set -x
zpool status )
fi
# TODO fix/re-enable
#[ `mount |grep /home |grep -c 'type zfs'` -ge 1 ] && failexit 109 "! Home already appears to be ON zfs!"
# bigvaiterazfsNB/home on /home type zfs (rw,noatime,xattr,noacl)
# Is /home a dir hanging off root or sep partn?
#sephome=`df /home |grep /home |awk '{ print $1 }'`
hmnt=$(mount |grep /home |awk '{ print $1 }' |head -n 1) # 1st line only
# bigvaiterazfsNB/home OR /dev/sdX9
roothome=0
homespc=0
if [ `echo $hmnt |grep -c '/dev'` -gt 0 ]; then
echo '';logecho "o Your /home appears to be on $hmnt"
df -h /home
elif [ -d /home ]; then
logecho "o Your /home does not appear to be on a separate partition, is a directory on the root filesystem"
echo "...Please wait while I determine how much space it is using..."
homespc=$(du -s -k /home |awk '{print $1}') # 431484266 /home
logecho $homespc
roothome=1
else
failexit 201 "! This fallthru should not happen, cannot determine /home!"
fi
# skip header line and grab 3rd field (Used)
#[ $debugg -gt 0 ] && homespc=16011904 # 16GB # TODO testing - only set this if there is nothing in /home
[ $homespc = 0 ] && homespc=$(df -k /home |tail -n +2 |awk '{ print $3 }')
let hsbytes=$homespc*1024
# REF: https://unix.stackexchange.com/questions/222121/how-to-remove-a-column-or-multiple-columns-from-file-using-shell-command
# get a list of long drive names with short; strip out blank lines and unnec fields
/bin/ls -go /dev/disk/by-id /dev/disk/by-path \
|egrep -v 'part|wwn|total |dev/' \
|awk 'NF>0' \
|awk '{$1=$2=$3=$4=$5=$6=""; print $0}' \
|column -t \
> $tmpfile1
echo '';echo "o These are the hard drives found on your system:"
# NOT an unnec use of cat - REF: https://unix.stackexchange.com/questions/16279/should-i-care-about-unnecessary-cats
cat $tmpfile1
echo ''
# did we get passed a disk or existing ZFS pool?
argg=$1
[ "$argg" = "" ] && failexit 199 "! Cannot proceed - pass at least a disk device name (long or short form) OR zfs pool to move /home to!"
usepool=""; usedisk=""
if [ `grep -c $argg $tmpfile1` -gt 0 ]; then
logecho "o You apparently want me to use this disk:"
bothforms=`grep $argg $tmpfile1`
echo "$bothforms"
getlongdisk=`grep $argg $tmpfile1 |awk '{ print $1 }' |head -n 1`
shortdisk=${bothforms##*/} # strip off all leading "/"
shortdev=/dev/$shortdisk
usedisk=$getlongdisk
echo ''; logecho "o Using long-form diskname: $usedisk - Short form: $shortdev"
echo "^^ If this is incorrect, then rerun this script and use a more specific device name!"
# test for cd = add all results (tmpusingcd)
ttlusecd=0 # TOTAL
#TMPusecd
tucd=`echo $argg |egrep -c 'sr0|sr1|scd0|scd1|cdrom|cdrw|dvdrw'`
let ttlusecd=$ttlusecd+$tucd
tucd=`echo $shortdisk |egrep -c 'sr0|sr1|scd0|scd1|cdrom|cdrw|dvdrw'`
let ttlusecd=$ttlusecd+$tucd
# [ `echo $argg |grep -c sr1` -gt 0 ] && failexit 401 "! I cant use a CDROM device, wiseguy!!"
[ $ttlusecd -gt 0 ] && failexit 401 "! I cant put /home on a CDROM device, wiseguy!! Try again with a hard drive!"
# test for existing filesystem on destination disk - especially if sda!
echo "...Checking blkid and zpools to see if the disk you specified is OK to use..."
[ `echo $hmnt |grep -c $shortdev` -gt 0 ] && failexit 32768 "! You CRAZY MANIAC - you cant re-use your existing home disk in-place for ZFS!!"
alreadyf=`blkid |grep -c $argg`
alreadyf2=`blkid |grep -c $shortdev`
let alreadyf=$alreadyf+$alreadyf2
#/dev/sde1: LABEL="zredpool2" UUID="17065421584496359800" UUID_SUB="1595728817173195411" TYPE="zfs_member" PARTLABEL="zfs"
#/dev/sda2: LABEL="xubuntu1404" UUID="103f019e-1275-4c27-a972-5b5d3874b863" TYPE="ext4" PARTUUID="b680669e-02"
# ISSUE - blkid is not always up to date, not detecting newly created test pools!
alreadyf2=`zpool status |grep -c $usedisk`
let alreadyf=$alreadyf+$alreadyf2
alreadyf2=`zpool status |grep -c $shortdisk`
let alreadyf=$alreadyf+$alreadyf2
alreadyf2=`zpool status |grep -c $argg`
let alreadyf=$alreadyf+$alreadyf2
# NOTE empty GPT label will not show on blkid!
[ $alreadyf -gt 0 ] && failexit 502 "! Disk is already formatted/IN USE and needs to either be blank or have an empty GPT label: $shortdev / $usedisk"
# Check disk capacity against existing
# fdisk -l /dev/sdb |grep Disk |grep -v 'identifier'
# 1 2 3 4 5
#Disk /dev/sdb: 1000 GB, 1000202273280 bytes
dcap=`fdisk -l $shortdev |grep Disk |grep -v 'identifier' |awk '{print $5}'`
[ $debugg -gt 0 ] && logecho "dcap: $dcap ^^ homespc: $hsbytes"
# comma-sep nums - REF: https://unix.stackexchange.com/questions/113795/add-thousands-separator-in-a-number
if [ $dcap -lt $hsbytes ]; then
dcapcma=`printf "%'d" $dcap`
hsbcma=`printf "%'d" $hsbytes`
logecho "! Disk capacity of $usedisk is less than home data usage!"
logecho "Home: $hsbcma"
logecho "Disk: $dcapcma"
failexit 999 "! Selected Disk capacity of $usedisk is less than home data usage - choose a larger disk or use a larger zpool!"
fi
################################# POINT OF NO RETURN - POSSIBLE DATA DESTRUCTION AFTER THIS!
fdisk -l $shortdev |tee -a $logfile 2>>$logfile
echo '';logecho "YOU ARE ABOUT TO DESTRUCTIVELY GPT LABEL DISK: $usedisk"
echo "ENTER ADMIN PASSWORD TO PROCEED OR ^C: "
read
( set -x
zpool labelclear $shortdev
parted -s $shortdev mklabel gpt
fdisk -l $shortdev |tee -a $logfile)
elif [ `zfs list -d0 |grep -c $argg` -gt 0 ]; then
logecho "o You apparently want me to use this pre-existing ZFS pool for /home:"
zfs list -d0 |grep $argg |head -n 1
usepool=$argg
zp=$argg # using pre-existing pool
else
failexit 404 "! Cannot proceed - $argg was not found on the system!"
fi # did we get passed a disk or existing ZFS pool?
# create the pool if needed
if [ "$usedisk" != "" ] && [ "$usepool" = "" ]; then
[ "$zp" = "" ] && zp=zhome
# set a default name
(set -x
zpool create -o ashift=12 -o autoexpand=on -o autoreplace=on \
-O atime=off -O compression=lz4 \
$zp \
$usedisk
zpool status |awk 'NF>0'
)
fi
# from now on, we are using pool!
(set -x
[ $debugg -gt 0 ] && zfs destroy $zp/home
zfs create -o sharesmb=off $zp/home )
zfs list -p $zp
# TODO check for zfs pool free space vs home use
zpcap=`zfs list -p $zp |awk '{ print $3 }' |tail -n +2` # skip header and get bytes
[ $debugg -gt 0 ] && logecho "zpcap: $zpcap ^^ homespc: $hsbytes"
if [ $zpcap -lt $hsbytes ]; then
zpcapcma=`printf "%'d" $zpcap`
hsbcma=`printf "%'d" $hsbytes`
logecho "! Usable ZFS pool capacity of $zp is less than home data usage!"
logecho "Home: $hsbcma"
logecho "Pool: $zpcapcma"
failexit 919 "! Selected ZFS pool $zp is smaller than home data usage - choose a larger disk or use a larger zpool!"
fi
# Permission was already given, but make sure it's OK to logoff all users
logecho "! NOTE: by proceeding from here, you will be shutting down the X window manager (GUI) and LOGGING OFF all non-root users!"
logecho ' /^^ MAKE SURE ALL YOUR DATA IS BACKED UP / SAVED BEFORE PROCEEDING ^^\'
logecho "You need to be DIRECTLY logged into tty1 or similar as the root userid!"
logecho "ENTER ADMIN PASSWORD TO PROCEED, OR ^C - you need to be running this script directly as root without using sudo!"
read
# Determine WM
xwm=`pstree -psu -A|grep Xorg`
# |-lightdm(1325)-+-Xorg(1388)
xwmedit=`echo $xwm |awk -F\( '{ print $1 }'`
#|-lightdm
xwmedit2=${xwmedit##*-} # strip off to "-" ${tmp2##*-}
#lightdm
[ "$xwmedit2" = "" ] || service $xwmedit2 stop
sleep 2
# OK so far, check if anything in /home is locked
function checkhomelock () {
flocks=`lsof |grep -c /home`
}
logecho "`date` ! Force-logging off anyone who is locking /home files..."
# xxxxx workaround, root getting fragged off
flocks=0
while [ $flocks -gt 0 ]; do
[ $debugg -gt 0 ] && fopts="-i "
# for myuser in `w -h |grep -v root |awk '{print $1}' |sort |uniq`; do
# fuser $fopts -u -v $myuser
fuser $fopts -k -u -v -m /home
# done
checkhomelock
sleep 5;date
done
lsof |grep /home
logecho "o All /home files should be free!"
du -s -h /home
logecho "`date` - Copying /home data over to /$zp/home"
cd /$zp/home || failexit 405 "! FARK - I cant cd to /$zp/home !"
cd /home; df -hT /home /$zp
# had problems with ownership permissions
#time tar cpf - * |pv |(cd /$zp/home; tar xpf - ) || failexit 1000 "! Copying home data failed - check free space!"
# xxxxx 2024.0404 EXPERIMENTAL but appears to work
time rsync -r -t -p -o -g -v --delete -l -s \
--exclude=.thumbnails/* \
/home/ \
/$zp/home \
2>~/rsync-error.log \
|| failexit 1000 "! Copying home data failed - check free space!"
date
df -hT |grep /home
if [ $debugg -gt 0 ]; then
logecho "PK to unmount old /home or move it out of the way:"
read
else
logecho "Unmounting old /home / moving it out of the way:"
fi
if [ $roothome -gt 0 ]; then
mv -v /home /home--old
ls -l / |grep home
else
cd
umount /home
fi
# SKIP edit fstab for noauto - NO, too dangerous to risk
zfs set mountpoint=/home $zp/home
df -h
[ "$xwmedit2" = "" ] || service $xwmedit2 start
logecho "`date` - Finished migrating /home"
zfs snapshot -r $zp@snapshot1
zfs list -r -t snapshot
logecho " -- Dont forget to restart the window manager if needed, and edit /etc/fstab - put /home as noauto!"
logecho "Example: # service lightdm restart"
logecho "EXAMPLE /etc/fstab:"
logecho "LABEL=home /home ext4 defaults,noauto,noatime,errors=remount-ro 0 1"
exit;
REQUIRES:
o Working 'zpool status' and 'zfs list'
o grep, awk, column
o parted, fdisk, blkid
o pstree, lsof, fuser
o tar, pv
lrwxrwxrwx 1 9 Apr 26 13:20 usb-VBOX_HARDDISK-0:0 -> ../../sde
lrwxrwxrwx 1 9 Apr 26 13:20 pci-0000:00:0b.0-usb-0:1:1.0-scsi-0:0:0:0 -> ../../sde
lrwxrwxrwx 1 9 Apr 26 13:20 pci-0000:00:14.0-scsi-0:0:0:0 -> ../../sdb
lrwxrwxrwx 1 9 Apr 26 13:20 pci-0000:00:16.0-sas-0x00060504030201a0-lun-0 -> ../../sdc
# column omits leading spaces :)
$ ls -go /dev/disk/by-id /dev/disk/by-path |egrep -v 'part|wwn|total |dev/' |awk 'NF>0' |awk '{$1=$2=$3=$4=$5=$6=""; print $0}'|column -t
ata-VBOX_CD-ROM_VB2-01700376 -> ../../sr0
ata-VBOX_HARDDISK_VB2cf5f3dc-6b93417b -> ../../sdg
ata-VBOX_HARDDISK_VB3c729abf-3f210fcb -> ../../sdd
ata-VBOX_HARDDISK_VB409c8b16-836d593c -> ../../sdf
ata-VBOX_HARDDISK_VBb85ec192-1f9a60c7 -> ../../sda
usb-VBOX_HARDDISK-0:0 -> ../../sde
pci-0000:00:0b.0-usb-0:1:1.0-scsi-0:0:0:0 -> ../../sde
pci-0000:00:14.0-scsi-0:0:0:0 -> ../../sdb
pci-0000:00:16.0-sas-0x00060504030201a0-lun-0 -> ../../sdc
alreadyf=`blkid |grep -c $argg`
#/dev/sde1: LABEL="zredpool2" UUID="17065421584496359800" UUID_SUB="1595728817173195411" TYPE="zfs_member" PARTLABEL="zfs" PARTUUID="05762e8f-a4e7-fe42-9b1b-3b2431b1f967"
#/dev/sde9: PARTUUID="9a508052-75aa-3047-a9ef-38d8c2d14649"
#/dev/sda2: LABEL="xubuntu1404" UUID="103f019e-1275-4c27-a972-5b5d3874b863" TYPE="ext4" PARTUUID="b680669e-02"
#/dev/sda3: LABEL="rootantiX-16" UUID="264d2bdc-d4df-4d76-a1dc-3096a5e68bb1" TYPE="ext4" PARTUUID="b680669e-03"
#/dev/sda1: PARTUUID="b680669e-01"
alreadyf2=`blkid |grep -c $shortdev`
let alreadyf=$alreadyf+$alreadyf2
#/dev/sde1: LABEL="zredpool2" UUID="17065421584496359800" UUID_SUB="1595728817173195411" TYPE="zfs_member" PARTLABEL="zfs"
#/dev/sda2: LABEL="xubuntu1404" UUID="103f019e-1275-4c27-a972-5b5d3874b863" TYPE="ext4" PARTUUID="b680669e-02"
#/dev/sda1: PARTUUID="b680669e-01"
Reference in New Issue View Git Blame Copy Permalink