274 lines
12 KiB
Markdown
274 lines
12 KiB
Markdown
+++
|
|
title = "Cluster layout management"
|
|
weight = 20
|
|
+++
|
|
|
|
The cluster layout in Garage is a table that assigns to each node a role in
|
|
the cluster. The role of a node in Garage can either be a storage node with
|
|
a certain capacity, or a gateway node that does not store data and is only
|
|
used as an API entry point for faster cluster access.
|
|
An introduction to building cluster layouts can be found in the [production deployment](@/documentation/cookbook/real-world.md) page.
|
|
|
|
In Garage, all of the data that can be stored in a given cluster is divided
|
|
into slices which we call *partitions*. Each partition is stored by
|
|
one or several nodes in the cluster
|
|
(see [`replication_factor`](@/documentation/reference-manual/configuration.md#replication_factor)).
|
|
The layout determines the correspondence between these partitions,
|
|
which exist on a logical level, and actual storage nodes.
|
|
|
|
## How cluster layouts work in Garage
|
|
|
|
A cluster layout is composed of the following components:
|
|
|
|
- a table of roles assigned to nodes, defined by the user
|
|
- an optimal assignation of partitions to nodes, computed by an algorithm that is ran once when calling `garage layout apply` or the ApplyClusterLayout API endpoint
|
|
- a version number
|
|
|
|
Garage nodes will always use the cluster layout with the highest version number.
|
|
|
|
Garage nodes also maintain and synchronize between them a set of proposed role
|
|
changes that haven't yet been applied. These changes will be applied (or
|
|
canceled) in the next version of the layout.
|
|
|
|
All operations on the layout can be realized using the `garage` CLI or using the
|
|
[administration API endpoint](@/documentation/reference-manual/admin-api.md).
|
|
We give here a description of CLI commands, the admin API semantics are very similar.
|
|
|
|
The following commands insert modifications to the set of proposed role changes
|
|
for the next layout version (but they do not create the new layout immediately):
|
|
|
|
```bash
|
|
garage layout assign [...]
|
|
garage layout remove [...]
|
|
```
|
|
|
|
The following command can be used to inspect the layout that is currently set in the cluster
|
|
and the changes proposed for the next layout version, if any:
|
|
|
|
```bash
|
|
garage layout show
|
|
```
|
|
|
|
The following commands create a new layout with the specified version number,
|
|
that either takes into account the proposed changes or cancels them:
|
|
|
|
```bash
|
|
garage layout apply --version <new_version_number>
|
|
garage layout revert --version <new_version_number>
|
|
```
|
|
|
|
The version number of the new layout to create must be 1 + the version number
|
|
of the previous layout that existed in the cluster. The `apply` and `revert`
|
|
commands will fail otherwise.
|
|
|
|
## Warnings about Garage cluster layout management
|
|
|
|
**⚠️ Never make several calls to `garage layout apply` or `garage layout
|
|
revert` with the same value of the `--version` flag. Doing so can lead to the
|
|
creation of several different layouts with the same version number, in which
|
|
case your Garage cluster will become inconsistent until fixed.** If a call to
|
|
`garage layout apply` or `garage layout revert` has failed and `garage layout
|
|
show` indicates that a new layout with the given version number has not been
|
|
set in the cluster, then it is fine to call the command again with the same
|
|
version number.
|
|
|
|
If you are using the `garage` CLI by typing individual commands in your
|
|
shell, you shouldn't have much issues as long as you run commands one after
|
|
the other and take care of checking the output of `garage layout show`
|
|
before applying any changes.
|
|
|
|
If you are using the `garage` CLI or the admin API to script layout changes,
|
|
follow the following recommendations:
|
|
|
|
- If using the CLI, make all of your `garage` CLI calls to the same RPC host.
|
|
If using the admin API, make all of your API calls to the same Garage node. Do
|
|
not connect to individual nodes to send them each a piece of the layout changes
|
|
you are making, as the changes propagate asynchronously between nodes and might
|
|
not all be taken into account at the time when the new layout is applied.
|
|
|
|
- **Only call `garage layout apply`/ApplyClusterLayout once**, and call it
|
|
**strictly after** all of the `layout assign` and `layout remove`
|
|
commands/UpdateClusterLayout API calls have returned.
|
|
|
|
|
|
## Understanding unexpected layout calculations
|
|
|
|
When adding, removing or modifying nodes in a cluster layout, sometimes
|
|
unexpected assignations of partitions to node can occur. These assignations
|
|
are in fact normal and logical, given the objectives of the algorithm. Indeed,
|
|
**the layout algorithm prioritizes moving less data between nodes over
|
|
achieving equal distribution of load. It also tries to use all links between
|
|
pairs of nodes in equal proportions when moving data.** This section presents
|
|
two examples and illustrates how one can control Garage's behavior to obtain
|
|
the desired results.
|
|
|
|
### Example 1
|
|
|
|
In this example, a cluster is originally composed of 3 nodes in 3 different
|
|
zones (data centers). The three nodes are of equal capacity, therefore they
|
|
are all fully exploited and all store a copy of all of the data in the cluster.
|
|
|
|
Then, a fourth node of the same size is added in the datacenter `dc1`.
|
|
As illustrated by the following, **Garage will by default not store any data on the new node**:
|
|
|
|
```
|
|
$ garage layout show
|
|
==== CURRENT CLUSTER LAYOUT ====
|
|
ID Tags Zone Capacity Usable capacity
|
|
b10c110e4e854e5a node1 dc1 1000.0 MB 1000.0 MB (100.0%)
|
|
a235ac7695e0c54d node2 dc2 1000.0 MB 1000.0 MB (100.0%)
|
|
62b218d848e86a64 node3 dc3 1000.0 MB 1000.0 MB (100.0%)
|
|
|
|
Zone redundancy: maximum
|
|
|
|
Current cluster layout version: 6
|
|
|
|
==== STAGED ROLE CHANGES ====
|
|
ID Tags Zone Capacity
|
|
a11c7cf18af29737 node4 dc1 1000.0 MB
|
|
|
|
|
|
==== NEW CLUSTER LAYOUT AFTER APPLYING CHANGES ====
|
|
ID Tags Zone Capacity Usable capacity
|
|
b10c110e4e854e5a node1 dc1 1000.0 MB 1000.0 MB (100.0%)
|
|
a11c7cf18af29737 node4 dc1 1000.0 MB 0 B (0.0%)
|
|
a235ac7695e0c54d node2 dc2 1000.0 MB 1000.0 MB (100.0%)
|
|
62b218d848e86a64 node3 dc3 1000.0 MB 1000.0 MB (100.0%)
|
|
|
|
Zone redundancy: maximum
|
|
|
|
==== COMPUTATION OF A NEW PARTITION ASSIGNATION ====
|
|
|
|
Partitions are replicated 3 times on at least 3 distinct zones.
|
|
|
|
Optimal partition size: 3.9 MB (3.9 MB in previous layout)
|
|
Usable capacity / total cluster capacity: 3.0 GB / 4.0 GB (75.0 %)
|
|
Effective capacity (replication factor 3): 1000.0 MB
|
|
|
|
A total of 0 new copies of partitions need to be transferred.
|
|
|
|
dc1 Tags Partitions Capacity Usable capacity
|
|
b10c110e4e854e5a node1 256 (0 new) 1000.0 MB 1000.0 MB (100.0%)
|
|
a11c7cf18af29737 node4 0 (0 new) 1000.0 MB 0 B (0.0%)
|
|
TOTAL 256 (256 unique) 2.0 GB 1000.0 MB (50.0%)
|
|
|
|
dc2 Tags Partitions Capacity Usable capacity
|
|
a235ac7695e0c54d node2 256 (0 new) 1000.0 MB 1000.0 MB (100.0%)
|
|
TOTAL 256 (256 unique) 1000.0 MB 1000.0 MB (100.0%)
|
|
|
|
dc3 Tags Partitions Capacity Usable capacity
|
|
62b218d848e86a64 node3 256 (0 new) 1000.0 MB 1000.0 MB (100.0%)
|
|
TOTAL 256 (256 unique) 1000.0 MB 1000.0 MB (100.0%)
|
|
```
|
|
|
|
While unexpected, this is logical because of the following facts:
|
|
|
|
- storing some data on the new node does not help increase the total quantity
|
|
of data that can be stored on the cluster, as the two other zones (`dc2` and
|
|
`dc3`) still need to store a full copy of everything, and their capacity is
|
|
still the same;
|
|
|
|
- there is therefore no need to move any data on the new node as this would be pointless;
|
|
|
|
- moving data to the new node has a cost which the algorithm decides to not pay if not necessary.
|
|
|
|
This distribution of data can however not be what the administrator wanted: if
|
|
they added a new node to `dc1`, it might be because the existing node is too
|
|
slow, and they wish to divide its load by half. In that case, what they need to
|
|
do to force Garage to distribute the data between the two nodes is to attribute
|
|
only half of the capacity to each node in `dc1` (in our example, 500M instead of 1G).
|
|
In that case, Garage would determine that to be able to store 1G in total, it
|
|
would need to store 500M on the old node and 500M on the added one.
|
|
|
|
|
|
### Example 2
|
|
|
|
The following example is a slightly different scenario, where `dc1` had two
|
|
nodes that were used at 50%, and `dc2` and `dc3` each have one node that is
|
|
100% used. All node capacities are the same.
|
|
|
|
Then, a node from `dc1` is moved into `dc3`. One could expect that the roles of
|
|
`dc1` and `dc3` would simply be swapped: the remaining node in `dc1` would be
|
|
used at 100%, and the two nodes now in `dc3` would be used at 50%. Instead,
|
|
this happens:
|
|
|
|
```
|
|
==== CURRENT CLUSTER LAYOUT ====
|
|
ID Tags Zone Capacity Usable capacity
|
|
b10c110e4e854e5a node1 dc1 1000.0 MB 500.0 MB (50.0%)
|
|
a11c7cf18af29737 node4 dc1 1000.0 MB 500.0 MB (50.0%)
|
|
a235ac7695e0c54d node2 dc2 1000.0 MB 1000.0 MB (100.0%)
|
|
62b218d848e86a64 node3 dc3 1000.0 MB 1000.0 MB (100.0%)
|
|
|
|
Zone redundancy: maximum
|
|
|
|
Current cluster layout version: 8
|
|
|
|
==== STAGED ROLE CHANGES ====
|
|
ID Tags Zone Capacity
|
|
a11c7cf18af29737 node4 dc3 1000.0 MB
|
|
|
|
|
|
==== NEW CLUSTER LAYOUT AFTER APPLYING CHANGES ====
|
|
ID Tags Zone Capacity Usable capacity
|
|
b10c110e4e854e5a node1 dc1 1000.0 MB 1000.0 MB (100.0%)
|
|
a235ac7695e0c54d node2 dc2 1000.0 MB 1000.0 MB (100.0%)
|
|
62b218d848e86a64 node3 dc3 1000.0 MB 753.9 MB (75.4%)
|
|
a11c7cf18af29737 node4 dc3 1000.0 MB 246.1 MB (24.6%)
|
|
|
|
Zone redundancy: maximum
|
|
|
|
==== COMPUTATION OF A NEW PARTITION ASSIGNATION ====
|
|
|
|
Partitions are replicated 3 times on at least 3 distinct zones.
|
|
|
|
Optimal partition size: 3.9 MB (3.9 MB in previous layout)
|
|
Usable capacity / total cluster capacity: 3.0 GB / 4.0 GB (75.0 %)
|
|
Effective capacity (replication factor 3): 1000.0 MB
|
|
|
|
A total of 128 new copies of partitions need to be transferred.
|
|
|
|
dc1 Tags Partitions Capacity Usable capacity
|
|
b10c110e4e854e5a node1 256 (128 new) 1000.0 MB 1000.0 MB (100.0%)
|
|
TOTAL 256 (256 unique) 1000.0 MB 1000.0 MB (100.0%)
|
|
|
|
dc2 Tags Partitions Capacity Usable capacity
|
|
a235ac7695e0c54d node2 256 (0 new) 1000.0 MB 1000.0 MB (100.0%)
|
|
TOTAL 256 (256 unique) 1000.0 MB 1000.0 MB (100.0%)
|
|
|
|
dc3 Tags Partitions Capacity Usable capacity
|
|
62b218d848e86a64 node3 193 (0 new) 1000.0 MB 753.9 MB (75.4%)
|
|
a11c7cf18af29737 node4 63 (0 new) 1000.0 MB 246.1 MB (24.6%)
|
|
TOTAL 256 (256 unique) 2.0 GB 1000.0 MB (50.0%)
|
|
```
|
|
|
|
As we can see, the node that was moved to `dc3` (node4) is only used at 25% (approximatively),
|
|
whereas the node that was already in `dc3` (node3) is used at 75%.
|
|
|
|
This can be explained by the following:
|
|
|
|
- node1 will now be the only node remaining in `dc1`, thus it has to store all
|
|
of the data in the cluster. Since it was storing only half of it before, it has
|
|
to retrieve the other half from other nodes in the cluster.
|
|
|
|
- The data which it does not have is entirely stored by the other node that was
|
|
in `dc1` and that is now in `dc3` (node4). There is also a copy of it on node2
|
|
and node3 since both these nodes have a copy of everything.
|
|
|
|
- node3 and node4 are the two nodes that will now be in a datacenter that is
|
|
under-utilized (`dc3`), this means that those are the two candidates from which
|
|
data can be removed to be moved to node1.
|
|
|
|
- Garage will move data in equal proportions from all possible sources, in this
|
|
case it means that it will tranfer 25% of the entire data set from node3 to
|
|
node1 and another 25% from node4 to node1.
|
|
|
|
This explains why node3 ends with 75% utilization (100% from before minus 25%
|
|
that is moved to node1), and node4 ends with 25% (50% from before minus 25%
|
|
that is moved to node1).
|
|
|
|
This illustrates the second principle of the layout computation: **if there is
|
|
a choice in moving data out of some nodes, then all links between pairs of
|
|
nodes are used in equal proportions** (this is approximately true, there is
|
|
randomness in the algorithm to achieve this so there might be some small
|
|
fluctuations, as we see above).
|