feat(XOSAN): add pics for disperse

This commit is contained in:
Olivier Lambert
2017-08-31 16:05:31 +02:00
parent 33b728fdc3
commit 848330c188
13 changed files with 15 additions and 9 deletions

BIN
docs/assets/disperse3.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 14 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 32 KiB

BIN
docs/assets/disperse5.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 17 KiB

BIN
docs/assets/disperse6.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 20 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 26 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 20 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 41 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 23 KiB

After

Width:  |  Height:  |  Size: 17 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 56 KiB

After

Width:  |  Height:  |  Size: 34 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 44 KiB

View File

@@ -20,7 +20,7 @@ Cons:
This is similar to **RAID5**: there is an [algorithm](https://en.wikipedia.org/wiki/Reed%E2%80%93Solomon_error_correction) that will generate a kind of parity, being able to continue to work even if 1 node is down. If you reintroduce the node, it will be "healed" automatically.
![picture disperse 3]()
![picture disperse 3](./assets/disperse3.png)
If you lose one node, your data are still here. This mode will give you **66% of your total disk space**.
@@ -28,7 +28,7 @@ If you lose one node, your data are still here. This mode will give you **66% of
Same than 3, like **RAID5**, you can lose 1 node without service interruption.
![picture disperse 5]()
![picture disperse 5](./assets/disperse5.png)
In this case, you'll be able to use up to **80%** of your total storage capacity!
@@ -36,7 +36,9 @@ In this case, you'll be able to use up to **80%** of your total storage capacity
It's very similar to **RAID6**. You can lose up to 2 nodes, it will continue to work in read and write.
![picture disperse 6]()
![picture disperse 6](./assets/disperse6.png)
![disperse 6 with 2 nodes down](./assets/disperse6_2nodesoff.png)
## Growing a disperse XOSAN
@@ -44,3 +46,5 @@ You can grow a replicated XOSAN by adding extra disperse volumes, in other words
* To grow a disperse 3, you need 3 new nodes. You'll add the total capacity of each disperse to make a distributed-disperse on 2x3 dispersed nodes.
* To grow a disperse 6, you need 6 new nodes.
![growing disperse](./assets/disperse3_grow.png)

View File

@@ -61,24 +61,24 @@ The easiest way is to use 3-way replication. You can lose completely 2 hosts, it
The usual deal is to create a "group" of 2 replicated nodes (2x2). In a picture:
![2x2 replication]()
![2x2 replication](./assets/replicate2x2.png)
#### 5 hosts
There is no way to use the local disks of 5 nodes in a replicated type. So you'll use 4 hosts in XOSAN, and the 5th would be also able to use the shared XOSAN SR, without participating directly to it.
![2x2 replication and 1 extra node]()
#### 6 hosts
You have 2 choices:
1. 2-way replication on 3 nodes (2x3)
2. 3-way replication on 2 nodes (3x2)
1. 2-way replication on 3 replicated nodes (2x3)
2. 3-way replication on 2 triplicated nodes (3x2)
There is more fault tolerance on mode 2, but less space usable. It's up to you!
![2x3 vs 3x2 modes]()
![2x3 mode](./assets/replicate2x3_full.png)
![3x2 mode](./assets/replicate3x2_full.png)
## Growing a replicated XOSAN

View File

@@ -37,3 +37,5 @@ For a 6 nodes setup, XOSAN will use in total:
* 12 vCPUs (usage is in general pretty low)
* 12 GiB RAM
* All Local LVM disk space
![picture disperse](./assets/disperse6.png)