fix incremental ecpool

Add --migrate feature: near-live migrate between clusters

README.md

@@ -2,7 +2,7 @@
 
 [![License](https://img.shields.io/github/license/EnterpriseVE/eve4pve-barc.svg)](https://www.gnu.org/licenses/gpl-3.0.en.html)
 
-Cross-Pool (live) Replication and near-live migration for Proxmox VE
+Cross-Pool asynchronous online-replication and near-live migration for Proxmox VE
 
 
 ```text
@@ -11,7 +11,7 @@ ______
 |   --|  _| . |_ -|_ -| . | | | -_|  _|
 |_____|_| |___|___|___|___|\_/|___|_|  
 
-Cross Pool (live) replication and near-live migration for Proxmox VE  
+Cross Pool asynchronous online-replication and near-live migration for Proxmox VE  
 
 Usage:
     crossover <COMMAND> [ARGS] [OPTIONS]
@@ -25,6 +25,7 @@ Commands:
     mirror               Replicate a stopped VM to another Cluster (full clone)
 
 Options:
+    --sshcipher          SSH Cipher to use for transfer (default: aes128-gcm@openssh.com,aes128-cbc)
     --vmid               The source+target ID of the VM, comma separated (eg. --vmid=100:100,101:101)
                          (The possibility to specify a different Target VMID is to not interfere with VMIDs on the
                          target cluster, or mark mirrored VMs on the destination)
@@ -86,10 +87,12 @@ It'll work according this scheme:
 * Retention policy: (eg. keep x snapshots on the source and y snapshots in the destination cluster)
 * Rewrites VM configurations so they match the new VMID and/or poolname on the destination
 * Secure an encrypted transfer (SSH), so it's safe to mirror between datacenter without an additional VPN
+* Near live-migrate: To move a VM from one Cluster to another, make an initial copy and re-run with --migrate. This will shutdown the VM on the source cluster and start it on the destination cluster.
 
 ## Installation of prerequisites
 
-```apt install git pv gawk jq
+```
+apt install git pv gawk jq curl
 
 ## Install the Script somewhere, eg to /opt
 
@@ -99,9 +102,9 @@ git clone https://github.com/lephisto/crossover/ /opt
 
 Ensure that you can freely ssh from the Node you plan to mirror _from_ to _all_ nodes in the destination cluster, as well as localhost.
 
-## Examples
+## Continuous replication between Clusters
 
-Mirror VM to another Cluster:
+Example 1: Mirror VM to another Cluster:
 
 ```
 root@pve01:~/crossover# ./crossover mirror --vmid=all --prefixid=99 --excludevmids=101 --destination=pve04 --pool=data2 --overwrite --online 
@@ -136,6 +139,7 @@ Full xmitted..........: 0 byte
 Differential Bytes ...: 372.96 KiB
 
 ```
+
 This example creates a mirror of VM 100 (in the source cluster) as VM 10100 (in the destination cluster) using the ceph pool "data2" for storing all attached disks. It will keep 4 Ceph snapshots prior the latest (in total 5) and 8 snapshots on the remote cluster. It will keep the VM on the target Cluster locked to avoid an accidental start (thus causing split brain issues), and will do it even if the source VM is running. 
 
 The use case is that you might want to keep a cold-standby copy of a certain VM on another Cluster. If you need to start it on the target cluster you just have to unlock it with `qm unlock VMID` there.
@@ -146,15 +150,87 @@ Another usecase could be that you want to migrate a VM from one cluster to anoth
 
 To minimize downtime and achive a near-live Migration from one Cluster to another it's recommended to do an initial Sync of a VM from the source to the destination cluster. After that, run the job again, and add the --migrate switch. This causes the source VM to be shut down prior snapshot + transfer, and be restarted on the destination cluster as soon as the incremental transfer is complete. Using --migrate will always try to start the VM on the destination cluster.
 
+Example 2: Near-live migrate VM from one cluster to another (Run initial replication first, which works online, then run with --migrate to shutdown on source, incrematally copy and start on destination):
+
+```
+root@pve01:~/crossover# ./crossover mirror --jobname=migrate --vmid=100 --destination=pve04 --pool=data2 --online
+ACTION: Onlinemirror
+Start mirror 2023-04-26 15:02:24
+VM 100 - Starting mirror for testubuntu
+VM 100 - Checking for VM 100 on destination cluster pve04 /etc/pve/nodes/*/qemu-server
+VM 100 - Transmitting Config for to destination pve04 VMID 100
+VM 100 - locked 100 [rc:0] on source
+VM 100 - locked 100 [rc:0] on destination
+VM 100 - Creating snapshot data/vm-100-disk-0@mirror-20230426150224
+VM 100 - Creating snapshot data/vm-100-disk-1@mirror-20230426150224
+VM 100 - unlocked source VM 100 [rc:0]
+VM 100 - F data/vm-100-disk-0@mirror-20230426150224: e:0:09:20 r:            c:[36.6MiB/s] a:[36.6MiB/s] 20.0GiB [===============================>] 100%
+VM 100 - created snapshot on 100 [rc:0]
+VM 100 - Disk Summary: Took 560 Seconds to transfer 20.00 GiB in a full run
+VM 100 - F data/vm-100-disk-1@mirror-20230426150224: e:0:00:40 r:            c:[50.7MiB/s] a:[50.7MiB/s] 2.00GiB [===============================>] 100%
+VM 100 - created snapshot on 100 [rc:0]
+VM 100 - Disk Summary: Took 40 Seconds to transfer 22.00 GiB in a full run
+VM 100 - Unlocking destination VM 100
+Finnished mirror 2023-04-26 15:13:47
+Job Summary: Bytes transferred 22.00 GiB for 2 Disks on 1 VMs in 00 hours 11 minutes 23 seconds
+VM Freeze OK/failed.......: 1/0
+RBD Snapshot OK/failed....: 2/0
+RBD export-full OK/failed.: 2/0
+RBD export-diff OK/failed.: 0/0
+Full xmitted..............: 22.00 GiB
+Differential Bytes .......: 0 Bytes
+
+root@pve01:~/crossover# ./crossover mirror --jobname=migrate --vmid=100 --destination=pve04 --pool=data2 --online --migrate
+ACTION: Onlinemirror
+Start mirror 2023-04-26 15:22:35
+VM 100 - Starting mirror for testubuntu
+VM 100 - Checking for VM 100 on destination cluster pve04 /etc/pve/nodes/*/qemu-server
+VM 100 - Migration requested, shutting down VM on pve01
+VM 100 - locked 100 [rc:0] on source
+VM 100 - locked 100 [rc:0] on destination
+VM 100 - Creating snapshot data/vm-100-disk-0@mirror-20230426152235
+VM 100 - Creating snapshot data/vm-100-disk-1@mirror-20230426152235
+VM 100 - I data/vm-100-disk-0@mirror-20230426152235: e:0:00:03 c:[1.29MiB/s] a:[1.29MiB/s] 4.38MiB
+VM 100 - Housekeeping: localhost data/vm-100-disk-0, keeping Snapshots for 0s
+VM 100 - Removing Snapshot localhost data/vm-100-disk-0@mirror-20230323162532 (2930293s) [rc:0]
+VM 100 - Removing Snapshot localhost data/vm-100-disk-0@mirror-20230426144911 (2076s) [rc:0]
+VM 100 - Removing Snapshot localhost data/vm-100-disk-0@mirror-20230426145632 (1637s) [rc:0]
+VM 100 - Removing Snapshot localhost data/vm-100-disk-0@mirror-20230426145859 (1492s) [rc:0]
+VM 100 - Removing Snapshot localhost data/vm-100-disk-0@mirror-20230426150224 (1290s) [rc:0]
+VM 100 - Housekeeping: pve04 data2/vm-100-disk-0-data, keeping Snapshots for 0s
+VM 100 - Removing Snapshot pve04 data2/vm-100-disk-0-data@mirror-20230426150224 (1293s) [rc:0]
+VM 100 - Disk Summary: Took 4 Seconds to transfer 4.37 MiB in a incremental run
+VM 100 - I data/vm-100-disk-1@mirror-20230426152235: e:0:00:00 c:[ 227 B/s] a:[ 227 B/s] 74.0 B
+VM 100 - Housekeeping: localhost data/vm-100-disk-1, keeping Snapshots for 0s
+VM 100 - Removing Snapshot localhost data/vm-100-disk-1@mirror-20230323162532 (2930315s) [rc:0]
+VM 100 - Removing Snapshot localhost data/vm-100-disk-1@mirror-20230426144911 (2098s) [rc:0]
+VM 100 - Removing Snapshot localhost data/vm-100-disk-1@mirror-20230426145632 (1659s) [rc:0]
+VM 100 - Removing Snapshot localhost data/vm-100-disk-1@mirror-20230426145859 (1513s) [rc:0]
+VM 100 - Removing Snapshot localhost data/vm-100-disk-1@mirror-20230426150224 (1310s) [rc:0]
+VM 100 - Housekeeping: pve04 data2/vm-100-disk-1-data, keeping Snapshots for 0s
+VM 100 - Removing Snapshot pve04 data2/vm-100-disk-1-data@mirror-20230426150224 (1313s) [rc:0]
+VM 100 - Disk Summary: Took 2 Seconds to transfer 4.37 MiB in a incremental run
+VM 100 - Unlocking destination VM 100
+VM 100 - Starting VM on pve01
+Finnished mirror 2023-04-26 15:24:25
+Job Summary: Bytes transferred 4.37 MiB for 2 Disks on 1 VMs in 00 hours 01 minutes 50 seconds
+VM Freeze OK/failed.......: 0/0
+RBD Snapshot OK/failed....: 2/0
+RBD export-full OK/failed.: 0/0
+RBD export-diff OK/failed.: 2/0
+Full xmitted..............: 0 Bytes
+Differential Bytes .......: 4.37 MiB
+```
+
 ## Things to check
 
-From Proxmox VE Hosts you want to backup you need to be able to ssh passwordless to all other Cluster hosts, that may hold VM's or Containers. This goes for the source and for the destination Cluster.
+From Proxmox VE Hosts you want to backup you need to be able to ssh passwordless to all other Cluster hosts, that may hold VM's or Containers. This goes for the source and for the destination Cluster. Doublecheck this.
 
 This is required for using the free/unfreeze and the lock/unlock function, which has to be called locally from that Host the guest is currently running on. Usually this works out of the box for the source cluster, but you may want to make sure that you can "ssh root@pvehost1...n" from every host to every other host in the cluster.
 
-For the Destination Cluster you need to copy your ssh-key to the first host in the cluster, and login once to every node
+For the Destination Cluster you need to copy your ssh-key to the first host in the cluster, and login once to every node in your cluster.
-in your cluster.
 
+Currently preflight checks don't include the check for enough resources in the destination cluster. Check beforehand that you don't exceed the maximum safe size of ceph in the destination cluster.
 
 ## Some words about Snapshot consistency and what qemu-guest-agent can do for you
 
@@ -297,3 +373,4 @@ Ceph Documentation:
 
 Proxmox Wiki:
 https://pve.proxmox.com/wiki/
+

crossover

@@ -1,5 +1,8 @@
 #!/bin/bash
 
+# Cross Pool Migration and incremental replication Tool for Proxmox VMs using Ceph.
+# Author: Bastian Mäuser <bma@netz.org>        
+
 LC_ALL="en_US.UTF-8"
 
 source rainbow.sh
@@ -13,13 +16,11 @@ declare opt_influx_jobname=''
 declare opt_influx_job_metrics='crossover_xmit'
 declare opt_influx_summary_metrics='crossover_jobs'
 
-# Cross Pool Migration and incremental replication Tool for Proxmox VMs using Ceph.
+name=$(basename "$0")
-# Author: Bastian Mäuser <bma@netz.org>        
+# readonly variables
-
+declare -r NAME=$name
-declare -r VERSION=0.7
+declare -r VERSION=0.8
-declare -r NAME=$(basename "$0")
 declare -r PROGNAME=${NAME%.*}
-
 declare -r PVE_DIR="/etc/pve"
 declare -r PVE_NODES="$PVE_DIR/nodes"
 declare -r QEMU='qemu-server'
@@ -27,14 +28,21 @@ declare -r QEMU_CONF_CLUSTER="$PVE_NODES/*/$QEMU"
 declare -r EXT_CONF='.conf'
 declare -r PVFORMAT_FULL='e:%t r:%e c:%r a:%a %b %p'
 declare -r PVFORMAT_SNAP='e:%t c:%r a:%a %b'
+logfile=$(mktemp)
+declare -r LOG_FILE=$logfile
 
-declare -r LOG_FILE=$(mktemp)
+# associative global arrays 
-
 declare -A -g pvnode
 declare -A -g dstpvnode
 declare -A -g svmids
 declare -A -g dvmids
 
+declare -g scluster
+declare -g dcluster
+declare -g scephversion
+declare -g dcephversion
+
+# global integers
 declare -g -i perf_freeze_ok=0
 declare -g -i perf_freeze_failed=0
 declare -g -i perf_ss_ok=0
@@ -51,11 +59,16 @@ declare -g -i perf_bytes_total=0
 declare -g -i perf_vm_running=0
 declare -g -i perf_vm_stopped=0
 declare -g -i perf_snaps_removed=0
+declare -g -i perf_vm_total=0
+declare -g -i perf_vm_ok=0
 
+# commandline parameters
 declare opt_destination
 declare opt_vm_ids=''
 declare opt_snapshot_prefix='mirror-'
 declare opt_rewrite=''
+declare opt_pool='rbd'
+declare opt_sshcipher='aes128-gcm@openssh.com,aes128-cbc'
 declare -i opt_prefix_id
 declare opt_exclude_vmids=''
 declare -i opt_debug=0
@@ -67,6 +80,7 @@ declare -i opt_keepdlock=0
 declare -i opt_overwrite=0
 declare -i opt_online=0
 declare -i opt_migrate=0
+declare -i opt_noconfirm=0
 declare opt_keep_local='0s'
 declare opt_keep_remote='0s'
 
@@ -75,6 +89,7 @@ declare -r redstconf='^\/etc\/pve\/nodes\/(.*)\/qemu-server\/([0-9]+).conf$'
 declare -r recephimg='([a-zA-Z0-9]+)\:(.*)'
 declare -r restripsnapshots='/^$/,$d'
 declare -r redateex='^([0-9]{4})([0-9]{2})([0-9]{2})([0-9]{2})([0-9]{2})([0-9]{2})$'
+declare -r restripansicolor='s/\x1b\[[0-9;]*m//g'
 
 function usage(){
    shift
@@ -104,11 +119,12 @@ Commands:
     mirror               Replicate a stopped VM to another Cluster (full clone)
 
 Options:
+    --sshcipher          SSH Cipher to use for transfer (default: aes128-gcm@openssh.com,aes128-cbc)
     --vmid               The source+target ID of the VM/CT, comma separated (eg. --vmid=100:100,101:101), or all for all
     --prefixid           Prefix for VMID's on target System [optional]
     --excludevmids       Exclusde VM IDs when using --vmid==all
     --destination        Target PVE Host in target pool. e.g. --destination=pve04
-    --pool               Ceph pool name in target pool. e.g. --pool=data
+    --pool               Ceph pool name in target pool. e.g. --pool=data [default=rbd]
     --keeplocal          How many additional Snapshots to keep locally. e.g. --keeplocal=2d
     --keepremote         How many additional Snapshots to keep remote. e.g. --keepremote=7d
     --rewrite            PCRE Regex to rewrite the Config Files (eg. --rewrite='s/(net0:)(.*)tag=([0-9]+)/\1\2tag=1/g' would
@@ -117,19 +133,20 @@ Options:
     --influxtoken        Influx API token with write permission
     --influxbucket       Influx Bucket to write to (e.g. --influxbucket=telegraf/autogen)
     --jobname            Descriptive name for the job, used in Statistics
+    --mail               Mail address to send report to, comma-seperated (e.g. --mail=admin@test.com,admin2@test.com)
 Switches:
     --online             Allow online Copy
-    --migrate            Stop VM on Source Cluster before final Transfer and restart on destination Cluster
+    --migrate            Stop VM on Source Cluster before final Transfer and start on destination Cluster
     --nolock             Don't lock source VM on Transfer (mainly for test purposes)
     --keep-slock         Keep source VM locked on Transfer
     --keep-dlock         Keep VM locked after transfer on Destination
     --overwrite          Overwrite Destination
+    --noconfirm          Don't ask for confirmation before starting --migrate mode (use with care!)
     --debug              Show Debug Output
 
 Report bugs to <mephisto@mephis.to>
 
 EOF
-    exit 1
 }
 
 function parse_opts(){
@@ -139,7 +156,7 @@ function parse_opts(){
     local args
     args=$(getopt \
            --options '' \
-           --longoptions=vmid:,prefixid:,excludevmids:,destination:,pool:,keeplocal:,keepremote:,rewrite:,influxurl:,influxorg:,influxtoken:,influxbucket:,jobname:,online,migrate,nolock,keep-slock,keep-dlock,overwrite,dry-run,debug \
+           --longoptions=sshcipher:,vmid:,prefixid:,excludevmids:,destination:,pool:,keeplocal:,keepremote:,rewrite:,influxurl:,influxorg:,influxtoken:,influxbucket:,jobname:,mail:,online,migrate,nolock,keep-slock,keep-dlock,overwrite,dry-run,noconfirm,debug,syslog \
            --name "$PROGNAME" \
            -- "$@") \
            || end_process 128
@@ -148,6 +165,7 @@ function parse_opts(){
 
     while true; do    
       case "$1" in
+        --sshcipher) opt_sshcipher=$2; shift 2;;
         --vmid) opt_vm_ids=$2; shift 2;;
         --prefixid) opt_prefix_id=$2; shift 2;;
         --excludevmids) opt_exclude_vmids=$2; shift 2;;
@@ -161,14 +179,17 @@ function parse_opts(){
         --influxtoken) opt_influx_token=$2; shift 2;;
         --influxbucket) opt_influx_bucket=$2; shift 2;;
         --jobname) opt_influx_jobname=$2; shift 2;;
+        --mail) opt_addr_mail="$2"; shift 2;;
         --online) opt_online=1; shift ;;
         --migrate) opt_migrate=1; shift ;;
         --dry-run) opt_dry_run=1; shift;;
+        --noconfirm) opt_noconfirm=1; shift;;
         --debug) opt_debug=1; shift;;
         --nolock) opt_lock=0; shift;;
         --keep-slock) opt_keepslock=1; shift;;
         --keep-dlock) opt_keepdlock=1; shift;;
         --overwrite) opt_overwrite=1; shift;;
+        --syslog) opt_syslog=1; shift;;
         --) shift; break;;
         *) break;;
       esac
@@ -181,7 +202,7 @@ function parse_opts(){
         log info "============================================"
         log info "Proxmox VE Version:"
 
-        echowhite $(pveversion)
+        echowhite "$(pveversion)"
 
         log info "============================================"
     fi
@@ -280,7 +301,14 @@ function exist_file(){
 function lookupcephpool() {
     pvehost=$1
     pvepoolname=$2
-    res=$(ssh $pvehost cat /etc/pve/storage.cfg | sed -n "/rbd: $pvepoolname/,/^$/p" | grep pool | cut -d " " -f 2)
+    res=$(ssh $pvehost cat /etc/pve/storage.cfg | sed -n "/rbd: $pvepoolname/,/^$/p" | grep -E "\s+pool\s" | cut -d " " -f 2)
+    echo $res
+}
+
+function lookupdatapool() {
+    pvehost=$1
+    pvepoolname=$2
+    res=$(ssh $pvehost cat /etc/pve/storage.cfg | sed -n "/rbd: $pvepoolname/,/^$/p" | grep -E "\s+data-pool\s" | cut -d " " -f 2)
     echo $res
 }
 
@@ -302,20 +330,6 @@ function get_vm_ids(){
     echo "$data"
 }
 
-function get_config_file(){
-    local file_config=''
-
-    if exist_file "$QEMU_CONF_CLUSTER/$vm_id$EXT_CONF"; then
-        file_config=$(ls $QEMU_CONF_CLUSTER/$vm_id$EXT_CONF)
-
-    else
-        log error "VM $vm_id - Unknown technology or VMID not found: $QEMU_CONF_CLUSTER/$vm_id$EXT_CONF"
-        end_process 128
-    fi 
-
-    echo "$file_config"
-}
-
 function get_disks_from_config(){
     local disks;
     local file_config=$1
@@ -327,7 +341,7 @@ function get_disks_from_config(){
                 [[ "$line" == "" ]] && break
                 echo "$line" 
             done < "$file_config" | \
-            grep -P '^(?:((?:virtio|ide|scsi|sata|mp)\d+)|rootfs): ' | \
+            grep -P '^(?:((?:efidisk|virtio|ide|scsi|sata|mp)\d+)|rootfs): ' | \
             grep -v -P 'cdrom|none' | \
             grep -v -P 'backup=0' | \
             awk '{ split($0,a,","); split(a[1],b," "); print b[2]}')
@@ -339,6 +353,7 @@ function log(){
     local level=$1
     shift 1
     local message=$*
+    local syslog_msg=''
 
     case $level in
         debug) 
@@ -350,28 +365,32 @@ function log(){
 
         info) 
             echo -e "$message"; 
-            echo -e "$message" >> "$LOG_FILE";
+            echo -e "$message" | sed -e 's/\x1b\[[0-9;]*m//g' >> "$LOG_FILE";
-            [ $opt_syslog -eq 1 ] && logger -t "$PROGNAME" "$message"
+            syslog_msg=$(echo -e "$message" | sed -e ${restripansicolor})
+            [ $opt_syslog -eq 1 ] && logger -t "$PROGNAME" "$syslog_msg"
             ;;
 
         warn)
-            echo -n $(echoyellow "WARNING: ")
+            echo -n "$(echoyellow 'WARNING: ')"
-            echo $(echowhite "$message") 1>&2
+            echowhite "$message" 1>&2
-            echo -e "$message" >> "$LOG_FILE";            
+            echo -e "$message" | sed -e ${restripansicolor} >> "$LOG_FILE";            
-            [ $opt_syslog -eq 1 ] && logger -t "$PROGNAME" -p daemon.warn "$message"
+            syslog_msg=$(echo -e "$message" | sed -e ${restripansicolor})
+            [ $opt_syslog -eq 1 ] && logger -t "$PROGNAME" -p daemon.warn "$syslog_msg"
             ;;
         
         error)
-            echo -n $(echored "ERROR: ")
+            echo -n "$(echored 'ERROR: ')"
-            echo $(echowhite "$message") 1>&2
+            echowhite "$message" 1>&2
-            echo -e "$message" >> "$LOG_FILE";            
+            echo -e "$message" | sed -e ${restripansicolor} >> "$LOG_FILE";            
-            [ $opt_syslog -eq 1 ] && logger -t "$PROGNAME" -p daemon.err "$message"
+            syslog_msg=$(echo -e "$message" | sed -e ${restripansicolor})
+            [ $opt_syslog -eq 1 ] && logger -t "$PROGNAME" -p daemon.err "$syslog_msg"
             ;;
 
         *)  
             echo "$message" 1>&2
-            echo -e "$message" >> "$LOG_FILE";            
+            echo -e "$message" | sed -e ${restripansicolor} >> "$LOG_FILE";            
-            [ $opt_syslog -eq 1 ] && logger -t "$PROGNAME" "$message"
+            syslog_msg=$(echo -e "$message" | sed -e ${restripansicolor})
+            [ $opt_syslog -eq 1 ] && logger -t "$PROGNAME" "$syslog_msg"
             ;;
     esac
 }
@@ -403,6 +422,9 @@ function mirror() {
     local -i endjob
     local -i vmcount=0
     local -i diskcount=0
+    local -i startdowntime
+    local -i enddowntime
+    local -i ga_ping
 
     local disp_perf_freeze_failed
     local disp_perf_ss_failed
@@ -413,6 +435,9 @@ function mirror() {
     log info "Start mirror $(date "+%F %T")"
     startjob=$(date +%s)
 
+    get_ceph_version
+    log info "Local Ceph Version: $scephversion, Remote Ceph Version: $dcephversion"
+
     #create pid file
     local pid_file="/var/run/$PROGNAME.pid"
     if [[ -e "$pid_file" ]]; then
@@ -427,32 +452,61 @@ function mirror() {
         end_process 1
     fi
 
+    scluster=$(grep cluster_name /etc/pve/corosync.conf | cut -d " " -f 4)
+    dcluster=$(ssh "$opt_destination" grep cluster_name /etc/pve/corosync.conf | cut -d " " -f 4)
+
+    if [ $opt_migrate -eq 1 ] && [ $opt_noconfirm -eq 0 ]; then
+        echo "VM(s) $opt_vm_ids will subsequently be shutdown on [$scluster] and started on [$dcluster]"
+        read -p "Do you want to proceed? (yes/no) " yn
+        case $yn in 
+            yes ) echo ok, we will proceed;;
+            no ) echo exiting...;
+                exit;;
+            * ) echo invalid response;
+                exit 1;;
+        esac
+    fi
+
 
 
     map_source_to_destination_vmid
     map_vmids_to_host
     map_vmids_to_dsthost "$opt_destination"
 
+    if [ "$(check_pool_exist "$opt_pool")" -eq 0 ]; then
+        log error "Preflight check: Destination RBD-Pool $opt_pool does not exist."
+        end_process 255
+    fi
+
     for vm_id in $svmids; do
+        file_config="$PVE_NODES/${pvnode[$vm_id]}/$QEMU/$vm_id.conf"
+        if ! exist_file "$file_config"; then
+            log error "VM $vm_id - Preflight check: VM $vm_id does not exist on source cluster [$scluster] - skipping to next VM."
+            continue
+        fi
+        ga_ping=$(gaping "$vm_id")
+        log debug "ga_ping: $ga_ping"
+        if [ "$ga_ping" -eq 255 ] ; then    #vm running but no qemu-guest-agent answering
+            log error "VM $vm_id - Preflight check: VM $vm_id on source cluster [$scluster] has no qemu-guest-agent running - skipping to next VM."
+            continue
+        fi
         (( vmcount++ ))
-        local file_config; file_config=$(get_config_file)
-        [ -z "$file_config" ] && continue
         local disk=''
         dvmid=${dvmids[$vm_id]}
         vmname=$(cat $PVE_NODES/"${pvnode[$vm_id]}"/$QEMU/"$vm_id".conf | sed -e ''$restripsnapshots'' | grep "name\:" | cut -d' ' -f 2)
         log info "VM $vm_id - Starting mirror for $(echowhite "$vmname")"
         srcvmgenid=$(cat $PVE_NODES/"${pvnode[$vm_id]}"/$QEMU/"$vm_id".conf | sed -e ''$restripsnapshots'' | grep vmgenid | sed -r -e 's/^vmgenid:\s(.*)/\1/')
-        dstvmgenid=$(ssh $opt_destination cat $PVE_NODES/"${dstpvnode[$dvmid]}"/$QEMU/"$dvmid".conf 2>/dev/null | grep vmgenid | sed -e ''$restripsnapshots'' | sed -r -e 's/^vmgenid:\s(.*)/\1/')
+        dstvmgenid=$(ssh "$opt_destination" cat $PVE_NODES/"${dstpvnode[$dvmid]}"/$QEMU/"$dvmid".conf 2>/dev/null | grep vmgenid | sed -e ''$restripsnapshots'' | sed -r -e 's/^vmgenid:\s(.*)/\1/')
-        log info "VM $vm_id - Checking for VM $dvmid on Destination Host $opt_destination $QEMU_CONF_CLUSTER"
+        log info "VM $vm_id - Checking for VM $dvmid on destination cluster $opt_destination $QEMU_CONF_CLUSTER"
         log debug "DVMID:$dvmid srcvmgenid:$srcvmgenid dstvmgenid:$dstvmgenid"
-        conf_on_destination=$(ssh $opt_destination "ls -d $QEMU_CONF_CLUSTER/$dvmid$EXT_CONF 2>/dev/null")
+        conf_on_destination=$(ssh "$opt_destination" "ls -d $QEMU_CONF_CLUSTER/$dvmid$EXT_CONF 2>/dev/null")
         [[ "$conf_on_destination" =~ $redstconf ]]
         host_on_destination=${BASH_REMATCH[1]}
 
         if [ $host_on_destination ]; then
             dststatus=$(ssh root@${dstpvnode[$dvmid]} qm status $dvmid|cut -d' ' -f 2)
             if [ $dststatus == "running" ]; then
-                log error "VM is running on Destination. bailing out"
+                log error "VM is running on Destination Cluster [$dcluster]. bailing out"
                 end_process 255
             fi
         fi
@@ -468,30 +522,35 @@ function mirror() {
                 log error "Source VM genid ($srcvmgenid) doesn't match destination VM genid ($dstvmgenid). This should not happen. Bailing out.."
                 end_process 255
             fi
-            log info "VM $vm_id - Transmitting Config for to destination $opt_destination VMID $dvmid"
+            log info "VM $vm_id - Transmitting Config for VM $vm_id to destination $opt_destination VMID $dvmid"
             rewriteconfig $PVE_NODES/"${pvnode[$vm_id]}"/$QEMU/"$vm_id".conf $opt_destination "$opt_pool" $PVE_NODES/"$opt_destination"/$QEMU/"$dvmid".conf "$dvmid"
             map_vmids_to_dsthost "$opt_destination"
         fi
 
-        #--move so we need to shutdown?
+        #--move so we need to shutdown and remove from ha group?
         if [ $opt_migrate -eq 1 ]; then
-            log info "VM $vm_id - Shutting down VM on ${pvnode[$vm_id]}"
+            log info "VM $vm_id - Migration requested, shutting down VM on ${pvnode[$vm_id]}"
-            do_run "ssh root@"${pvnode[$vm_id]}" qm shutdown "$vm_id >/dev/null
+            if [ "$(get_ha_status "$vm_id")" == "started" ]; then
+                log info "VM $vm_id - remove from HA"
+                do_run "ha-manager remove $vm_id"
+            fi
+            do_run "ssh root@${pvnode[$vm_id]} qm shutdown $vm_id >/dev/null"
+            startdowntime=$(date +%s)
         fi
 
         #Lock on source + destination
         if [ $opt_lock -eq 1 ]; then
-            do_run "ssh root@"${pvnode[$vm_id]}" qm set "$vm_id" --lock backup" >/dev/null
+            do_run "ssh root@""${pvnode[$vm_id]}"" qm set ""$vm_id"" --lock backup" >/dev/null
-            log info "VM $vm_id - locked $vm_id [rc:$?]"
+            log info "VM $vm_id - locked $vm_id [rc:$?] on source"
-            do_run "ssh root@"${dstpvnode[$dvmid]}" qm set "$dvmid" --lock backup" >/dev/null
+            do_run "ssh root@""${dstpvnode[$dvmid]}"" qm set ""$dvmid"" --lock backup" >/dev/null
-            log info "VM $dvmid - locked $dvmid [rc:$?]"
+            log info "VM $dvmid - locked $dvmid [rc:$?] on destination"
         fi
-        #Freeze fs only if no migration running
+        #Freeze fs only if no migration running and qemu-guest-agent okay.
-        if [ $opt_migrate -eq 0 ]; then
+        if [ $opt_migrate -eq 0 ] && [ $ga_ping -eq 0 ]; then
             vm_freeze "$vm_id" "${pvnode[$vm_id]}" >/dev/null
             freezerc=$?
             if [ $freezerc -gt 0 ]; then
-                log error "VM $vm_id - QEMU-Guest could not fsfreeze on guest."
+                log warn "VM $vm_id - QEMU-Guest could not fsfreeze on guest."
                 (( perf_freeze_failed++ ))
             else
                 (( perf_freeze_ok++ ))
@@ -502,7 +561,7 @@ function mirror() {
             create_snapshot "$src_image_spec@$opt_snapshot_prefix$timestamp" 2>/dev/null
             ssrc=$?
             if [ $ssrc -gt 0 ]; then
-                log error "VM $vm_id - rbd snap failed."
+                log warn "VM $vm_id - rbd snap failed."
                 (( perf_ss_failed++ ))
             else
                 (( perf_ss_ok++ ))
@@ -533,6 +592,10 @@ function mirror() {
             [[ $dst_image_spec =~ ^.*\/(.*)$ ]]
             dst_image_name=${BASH_REMATCH[1]}-$src_image_pool_pve
             dst_image_pool=$(lookupcephpool $opt_destination $opt_pool)
+            dst_data_pool=$(lookupdatapool $opt_destination $opt_pool)
+            if [ -n "$dst_data_pool" ]; then
+                dst_data_opt="--data-pool $dst_data_pool"
+            fi
             snapshot_name="@$opt_snapshot_prefix$timestamp"
             localsnapcount=$(rbd ls -l $src_image_pool | grep $src_image_name@$opt_snapshot_prefix | cut -d ' ' -f 1|wc -l)
             if [ $localsnapcount -ge 2 ]; then
@@ -556,7 +619,7 @@ function mirror() {
                 #snapts=$(echo $currentlocal | sed -r -e 's/.*@mirror-(.*)/\1/')
                 snapshotsize=$(rbd du --pretty-format --format json $src_image_pool/$src_image_name|jq '.images[] | select (.snapshot_id == null) | {provisioned_size}.provisioned_size'|tail -1)
                 log debug "snapsize: $snapshotsize "
-                xmitjob="rbd export --rbd-concurrent-management-ops 8 $src_image_pool/$src_image_name$snapshot_name --no-progress - | tee >({ wc -c; } >/tmp/$PROGNAME.$pid.$dst_image_pool-$dst_image_name.size) | pv -s $snapshotsize -F \"VM $vm_id - F $src_image_pool/$src_image_name$snapshot_name: $PVFORMAT_FULL\" | ssh $opt_destination rbd import --image-format 2 - $dst_image_pool/$dst_image_name  2>/dev/null"
+                xmitjob="rbd export --rbd-concurrent-management-ops 8 $src_image_pool/$src_image_name$snapshot_name --no-progress - | tee >({ wc -c; } >/tmp/$PROGNAME.$pid.$dst_image_pool-$dst_image_name.size) | pv -s $snapshotsize -F \"VM $vm_id - F $src_image_pool/$src_image_name$snapshot_name: $PVFORMAT_FULL\" | ssh -c $opt_sshcipher $opt_destination rbd import --image-format 2 - $dst_image_pool/$dst_image_name $dst_data_opt 2>/dev/null"
                 # create initial snapshot on destination
                 log debug "xmitjob: $xmitjob"
                 startdisk=$(date +%s)
@@ -584,7 +647,7 @@ function mirror() {
                     #disk was not attached, or really nothing has changed..
                     snapshotsize=0
                 fi
-                xmitjob="rbd export-diff --no-progress --from-snap $opt_snapshot_prefix$basets $src_image_pool/$currentlocal - | tee >({ wc -c; } >/tmp/$PROGNAME.$pid.$dst_image_pool-$dst_image_name.size) | pv -F \"VM $vm_id - I $src_image_pool/$src_image_name$snapshot_name: $PVFORMAT_SNAP\" | ssh $opt_destination rbd import-diff --no-progress - $dst_image_pool/$dst_image_name"
+                xmitjob="rbd export-diff --no-progress --from-snap $opt_snapshot_prefix$basets $src_image_pool/$currentlocal - | tee >({ wc -c; } >/tmp/$PROGNAME.$pid.$dst_image_pool-$dst_image_name.size) | pv -F \"VM $vm_id - I $src_image_pool/$src_image_name$snapshot_name: $PVFORMAT_SNAP\" | ssh -c $opt_sshcipher $opt_destination rbd import-diff --no-progress - $dst_image_pool/$dst_image_name"
                 log debug "xmitjob: $xmitjob"
                 startdisk=$(date +%s)
                 do_run "$xmitjob"
@@ -619,8 +682,10 @@ function mirror() {
         fi
         #--migrate so start on destination?
         if [ $opt_migrate -eq 1 ]; then
-            log info "VM $dvmid - Starting VM on ${pvnode[$vm_id]}"
+            log info "VM $dvmid - Starting VM on node ${dstpvnode[$dvmid]} in cluster [$dcluster]"
-            do_run "ssh root@"${dstpvnode[$dvmid]}" qm start "$dvmid >/dev/null
+            do_run "ssh root@""${dstpvnode[$dvmid]}"" qm start "$dvmid >/dev/null
+            enddowntime=$(date +%s)
+            log info "VM $dvmid - Downtime: $(( enddowntime - startdowntime )) Seconds"
         fi
 
     done
@@ -645,6 +710,8 @@ function mirror() {
         cmd="curl --request POST \"$opt_influx_api_url/v2/write?org=$opt_influx_api_org&bucket=$opt_influx_bucket&precision=ns\" --header \"Authorization: Token $opt_influx_token\" --header \"Content-Type: text/plain; charset=utf-8\" --header \"Accept: application/json\" --data-binary '$influxlp'"
         do_run "$cmd"
     fi
+    (( perf_vm_ok++ ))
+    end_process 0
 }
 
 function do_housekeeping(){
@@ -686,6 +753,15 @@ function do_housekeeping(){
     done
 }
 
+function gaping() {
+    local vmid=$1
+    local rc
+    cmd="ssh root@${pvnode[$vmid]} qm guest cmd $vmid ping >/dev/null 2>&1"
+    eval "$cmd"
+    rc=$?
+    echo $rc
+}
+
 function create_snapshot(){
     local snap="$1"
     log info "VM $vm_id - Creating snapshot $snap"
@@ -701,8 +777,8 @@ function vm_freeze() {
     status=$(ssh root@"$fhost" qm status "$fvm"|cut -d' ' -f 2)
     if ! [[ "$status" == "running" ]]; then
         log info "VM $fvm - Not running, skipping fsfreeze-freeze"
-        return
         (( perf_vm_stopped++ ))
+        return
     else 
         (( perf_vm_running++ ))
     fi
@@ -742,7 +818,7 @@ function rewriteconfig(){
     else
         sedcmd='sed -e /^$/,$d'
     fi
-    cat "$oldconfig" | sed -r -e "s/^(virtio|ide|scsi|sata|mp)([0-9]+):\s([a-zA-Z0-9]+):(.*)-([0-9]+)-disk-([0-9]+).*,(.*)$/\1\2: $newpool:\4-$newvmid-disk-\6-\3,\7/g" | $sedcmd | sed -e '/^$/,$d' | grep -v "^parent:\s.*$" | ssh "$dst" "cat - >$newconfig"
+    cat "$oldconfig" | sed -r -e "s/^(efidisk|virtio|ide|scsi|sata|mp)([0-9]+):\s([a-zA-Z0-9]+):(.*)-([0-9]+)-disk-([0-9]+).*,(.*)$/\1\2: $newpool:\4-$newvmid-disk-\6-\3,\7/g" | $sedcmd | sed -e '/^$/,$d' | sed -e '/ide[0-9]:.*-cloudinit,media=cdrom.*/d' | grep -v "^parent:\s.*$" | ssh "$dst" "cat - >$newconfig"
 }
 
 function checkvmid(){
@@ -773,23 +849,25 @@ function do_run(){
 
 function end_process(){
     local -i rc=$1;
-#    if ! [[ -z "$startts" && -z "$endts" ]]; then
+    local -i runtime
-#        local -i runtime=$(expr $endts - $startts)
+    local -i bps
-#        local -i bps=$(expr $bytecount/$runtime)
+    local -i ss_total
-#    fi
+    local subject
-#    local subject="Ceph [VM:$vmok/$vmtotal SS:$snapshotok/$snapshottotal EX:$exportok/$exporttotal] [$(bytesToHuman "$bytecount")@$(bytesToHuman "$bps")/s]"
+    if ! [[ -z "$startjob" || -z "$endjob" ]]; then
-#    [ $rc != 0 ] && subject="$subject [ERROR]"
+        runtime=$(expr $endjob - $startjob)
+        bps=$(expr $perf_bytes_total/$runtime)
+    fi
+    ss_total=$(expr $perf_ss_ok + $perf_ss_failed)
+    subject="Crossover [VM:$perf_vm_ok/$vmcount SS:$perf_ss_ok/$ss_total]"
+    [ $rc != 0 ] && subject="[ERROR] $subject" || subject="[OK] $subject"
 
-    #send email
+    local mail;
-#    local mail;
+    for mail in $(echo "$opt_addr_mail" | tr "," "\n"); do
-#    local mailhead="Backup $imgcount Images in $vmcount VMs (Bytes: $bytecount)"
+        do_run "cat '$LOG_FILE' | mail -s '$subject' '$mail'"
-#    for mail in $(echo "$opt_addr_mail" | tr "," "\n"); do
+    done
-#        do_run "cat '$LOG_FILE' | mail -s '$subject' '$mail'"
-#    done
 
     #remove log
-#    rm "$LOG_FILE"
+    rm "$LOG_FILE"
-
     exit "$rc";
 }
 
@@ -806,6 +884,29 @@ function get_image_spec(){
     echo "$image_spec"
 }
 
+function get_ha_status() {
+    local havmid="$1"
+    ha_status=$(ha-manager status| grep vm:"$havmid" | cut -d " " -f 4| sed 's/.$//')
+    echo "$ha_status"
+}
+
+function check_pool_exist() {
+    local poolname="$1"
+    local -i exists=255
+    pool_status=$(ssh $opt_destination pvesm status|grep rbd|cut -d " " -f 1|grep $poolname)
+    if [ "$pool_status" == "$poolname" ]; then
+        exists=1
+    else
+        exists=0
+    fi
+    echo $exists
+}
+
+function get_ceph_version() {
+    scephversion=$(ceph -v | cut -d " " -f 3)
+    dcephversion=$(ssh $opt_destination ceph -v | cut -d " " -f 3)
+}
+
 function main(){    
     [ $# = 0 ] && usage;