The pvesr command line tool manages the {PVE} storage replication
framework. Storage replication brings redundancy for guests using
local storage and reduces migration time.
It replicates guest volumes to another node so that all data is available without using shared storage. Replication uses snapshots to minimize traffic sent over the network. Therefore, new data is sent only incrementally after an initial full sync. In the case of a node failure, your guest data is still available on the replicated node.
The replication will be done automatically in configurable intervals.
The minimum replication interval is one minute and the maximal interval is
once a week. The format used to specify those intervals is a subset of
systemd calendar events, see
Schedule Format section:
Every guest can be replicated to multiple target nodes, but a guest cannot get replicated twice to the same target node.
Each replications bandwidth can be limited, to avoid overloading a storage or server.
Virtual guest with active replication cannot currently use online migration.
Offline migration is supported in general. If you migrate to a node where
the guests data is already replicated only the changes since the last
synchronisation (so called delta) must be sent, this reduces the required
time significantly. In this case the replication direction will also switch
nodes automatically after the migration finished.
For example: VM100 is currently on nodeA and gets replicated to nodeB.
You migrate it to nodeB, so now it gets automatically replicated back from
nodeB to nodeA.
If you migrate to a node where the guest is not replicated, the whole disk data must send over. After the migration the replication job continues to replicate this guest to the configured nodes.
|
Important
|
High-Availability is allowed in combination with storage replication, but it has the following implications:
|
| Description | PVE type | Snapshots | Stable |
|---|---|---|---|
ZFS (local) |
zfspool |
yes |
yes |
{pve} has a very flexible replication scheduler. It is based on the systemd time calendar event format.[1] Calendar events may be used to refer to one or more points in time in a single expression.
Such a calendar event uses the following format:
[day(s)] [[start-time(s)][/repetition-time(s)]]
This allows you to configure a set of days on which the job should run.
You can also set one or more start times, it tells the replication scheduler
the moments in time when a job should start.
With this information we could create a job which runs every workday at 10
PM: 'mon,tue,wed,thu,fri 22' which could be abbreviated to: 'mon..fri
22', most reasonable schedules can be written quite intuitive this way.
|
Note
|
Hours are set in 24h format. |
To allow easier and shorter configuration one or more repetition times can
be set. They indicate that on the start-time(s) itself and the start-time(s)
plus all multiples of the repetition value replications will be done. If
you want to start replication at 8 AM and repeat it every 15 minutes until
9 AM you would use: '8:00/15'
Here you see also that if no hour separation (:) is used the value gets
interpreted as minute. If such a separation is used the value on the left
denotes the hour(s) and the value on the right denotes the minute(s).
Further, you can use * to match all possible values.
To get additional ideas look at more Examples below.
- days
-
Days are specified with an abbreviated English version:
sun, mon, tue, wed, thu, fri and sat. You may use multiple days as a comma-separated list. A range of days can also be set by specifying the start and end day separated by “..”, for examplemon..fri. Those formats can be also mixed. If omitted'*'is assumed. - time-format
-
A time format consists of hours and minutes interval lists. Hours and minutes are separated by
':'. Both, hour and minute, can be list and ranges of values, using the same format as days. First come hours then minutes, hours can be omitted if not needed, in this case'*'is assumed for the value of hours. The valid range for values is0-23for hours and0-59for minutes.
| Schedule String | Alternative | Meaning |
|---|---|---|
mon,tue,wed,thu,fri |
mon..fri |
Every working day at 0:00 |
sat,sun |
sat..sun |
Only on weekends at 0:00 |
mon,wed,fri |
— |
Only on Monday, Wednesday and Friday at 0:00 |
12:05 |
12:05 |
Every day at 12:05 PM |
*/5 |
0/5 |
Every five minutes |
mon..wed 30/10 |
mon,tue,wed 30/10 |
Monday, Tuesday, Wednesday 30, 40 and 50 minutes after every full hour |
mon..fri 8..17,22:0/15 |
— |
Every working day every 15 minutes between 8 AM and 6 PM and between 10 PM and 11 PM |
fri 12..13:5/20 |
fri 12,13:5/20 |
Friday at 12:05, 12:25, 12:45, 13:05, 13:25 and 13:45 |
12,14,16,18,20,22:5 |
12/2:5 |
Every day starting at 12:05 until 22:05, every 2 hours |
* |
*/1 |
Every minute (minimum interval) |
If a replication job encounters problems it will be placed in error state. In this state the configured replication intervals get suspended temporarily. Then we retry the failed replication in a 30 minute interval, once this succeeds the original schedule gets activated again.
This represents only the most common issues possible, depending on your setup there may be also another cause.
-
Network is not working.
-
No free space left on the replication target storage.
-
Storage with same storage ID available on target node
|
Note
|
You can always use the replication log to get hints about a problems cause. |
In the case of a grave error a virtual guest may get stuck on a failed node. You then need to move it manually to a working node again.
Lets assume that you have two guests (VM 100 and CT 200) running on node A and replicate to node B. Node A failed and can not get back online. Now you have to migrate the guest to Node B manually.
-
connect to node B over ssh or open its shell via the WebUI
-
check if that the cluster is quorate
# pvecm status
-
If you have no quorum we strongly advise to fix this first and make the node operable again. Only if this is not possible at the moment you may use the following command to enforce quorum on the current node:
# pvecm expected 1
|
Warning
|
If expected votes are set avoid changes which affect the cluster (for example adding/removing nodes, storages, virtual guests) at all costs. Only use it to get vital guests up and running again or to resolve the quorum issue itself. |
-
move both guest configuration files form the origin node A to node B:
# mv /etc/pve/nodes/A/qemu-server/100.conf /etc/pve/nodes/B/qemu-server/100.conf # mv /etc/pve/nodes/A/lxc/200.conf /etc/pve/nodes/B/lxc/200.conf
-
Now you can start the guests again:
# qm start 100 # pct start 200
Remember to replace the VMIDs and node names with your respective values.
You can use the web GUI to create, modify and remove replication jobs
easily. Additionally the command line interface (CLI) tool pvesr can be
used to do this.
You can find the replication panel on all levels (datacenter, node, virtual guest) in the web GUI. They differ in what jobs get shown: all, only node specific or only guest specific jobs.
Once adding a new job you need to specify the virtual guest (if not already
selected) and the target node. The replication
schedule can be set if the default of all
15 minutes is not desired. You may also impose rate limiting on a
replication job, this can help to keep the storage load acceptable.
A replication job is identified by an cluster-wide unique ID. This ID is composed of the VMID in addition to an job number. This ID must only be specified manually if the CLI tool is used.
Create a replication job which will run every 5 minutes with limited bandwidth of 10 mbps (megabytes per second) for the guest with guest ID 100.
# pvesr create-local-job 100-0 pve1 --schedule "*/5" --rate 10
Disable an active job with ID 100-0
# pvesr disable 100-0
Enable a deactivated job with ID 100-0
# pvesr enable 100-0
Change the schedule interval of the job with ID 100-0 to once a hour
# pvesr update 100-0 --schedule '*/00'
man 7 systemd.time for more information