Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

doc(XCP-ng): Adding multipathing guideline #316

Open
wants to merge 10 commits into
base: master
Choose a base branch
from
260 changes: 260 additions & 0 deletions docs/storage/multipathing.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,260 @@
---
sidebar_position: 1
---

# Multipathing

How to properly setup storage multipathing with Xen Orchestra and XCP-ng.

## iSCSI

### Requirements
* Two different network cards (It can be the same model)
* Dedicated network interfaces without VLAN tagging on XCP-ng host and storage unit
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this is also not necessary, we have a ton of users and customers running iSCSI storage over vlans, there's nothing wrong with it at all

* Two different switches (not stacked)
* Two VLANs **without L3 routing**
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

storage vlans can be routed, it works, it just takes a little more setup. Again I think we should remove these out of scope reqs

* Two IPv4 subnets **without L3 routing**
* Multiple targets per LUN on your storage unit
* iSCSI target ports are operating in portal mode

:::info
It is recommended to configure the network interfaces of XCP-ng host servers, switch interfaces, and storage array interfaces with Jumbo frames (MTU 9000).
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

All network interfaces, or those dedicated to storage?

I often heard @Fohdeesha say that MTU 9000 is usually not required nowadays, and complain about users who automatically put MTU 9000, not always doing it right, and causing all kinds of networking issues (I can remember a recent nightmarish ticket).

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please read the entire note.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So, I had read the entire note. But it doesn't answer my question. So I'll try to put it in a different way: you wrote the network interfaces of XCP-ng host servers, as if they were necessarily all included in a multipath setup. Is that always the case? What about the management interface? The live migration network when there's one? Networks dedicated to VM transfers? Should we be more specific regarding what network interfaces of the host must be configured with jumbo frames?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm only talking about iSCSI interfaces, not the others.

This is iSCSI/FC multipathing documentation: all other interfaces are out of scope and will never be mentioned.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I really really really advise against broadly recommending jumbo frames - on hardware from the last ten or even 15 years, on 10gbE interfaces there is no performance gain - and it opens up a huge box of potential issues that users are going to get themselves into. Something more specific like "if your storage vendor recommends jumbo frames, then do xxx" would be much safer

Copy link
Member

@stormi stormi Feb 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm only talking about iSCSI interfaces, not the others.

This is iSCSI/FC multipathing documentation: all other interfaces are out of scope and will never be mentioned.

I thought we had an agreement about this on thursday, after one full hour trying to make you consider my request, and now we're back again to "my document is good as is it, I know my stuff and I won't change it". That's not how peer reviewing works. We have a lot of experience in anticipating how users may interpret stuff the wrong way, and we've spent a lot more time arguing than it would have taken you to add the one or two words that would specify which NICs this applies to.

Your document exists in a broader context, a hypervisor with a complex networking setup, that it's easy to get wrong (ask @Fohdeesha about it). What's obvious to you ("all other interfaces are out of scope and will never be mentioned") is not necessarily obvious to users. When you have the opportunity to make this more obvious, always do. Let's not leave any ambiguity.

Jon's comment also raises important questions, and I kind of agree with him, but I won't take a decision about it for now as it probably requires more internal discussion.


If your switch cannot handle Jumbo frames, you can leave the default values (MTU 1500). But make sure that all elements are set to MTU 1500.
:::

:::warning
1. Make sure you do not use bond-type network interfaces on the host and on the storage unit for iSCSI interfaces.
2. Make sure you do not configure network routes on iSCSI interfaces.
:::

### Target architecture

#### Configuration example
| Path | Vlan | Subnet | Host PIF address | Storage Controller 1 address | Storage Controller 2 address |
| :---: | :---: | :---: | :---: | :---: | :---: |
| 🔵 | 421 | 10.42.1.0/24 | 10.42.1.11 | 10.42.1.101 | 10.42.1.102 |
| 🟢 | 422 | 10.42.2.0/24 | 10.42.2.11 | 10.42.2.101 | 10.42.2.102 |

#### Target architecture diagram
```mermaid
---
config:
look: handDrawn
theme: default
---
flowchart LR
classDef blueNode fill:#A7C8F0,stroke:#2C6693,color:#000000;
classDef greenNode fill:#A8E6A2,stroke:#3E8E41,color:#000000;
classDef grayNode fill:#D6D6D6,stroke:#8C8C8C,color:#000000;

subgraph server[XCP-ng host]
direction LR
subgraph Card2[Network card 2]
direction RL
card2pif1[pif4]
card2pif2[pif3]
end
subgraph Card1[Network card 1]
direction RL
card1pif1[pif1]
card1pif2[pif2]
end
end
subgraph Switch 2
vlan422[vlan422]:::greenNode
end
subgraph Switch 1
vlan421[vlan421]:::blueNode
end
subgraph storage[Storage unit]
direction LR
subgraph ctrl2[Controller 2]
direction RL
ctrl2-port1[port1]
ctrl2-port2[port2]
end
subgraph ctrl1[Controller 1]
direction RL
ctrl1-port1[port1]
ctrl1-port2[port2]
end
lun1@{ shape: lin-cyl, label: "LUN" }

end

card1-pif1-ip([10.42.1.11/24]):::blueNode
card1-pif2-ip([other networks]):::grayNode
card2-pif2-ip([10.42.2.11/24]):::greenNode
card2-pif1-ip([other networks]):::grayNode
ctrl1-port1-ip([10.42.1.101/24]):::blueNode
ctrl1-port2-ip([10.42.2.101/24]):::greenNode
ctrl2-port1-ip([10.42.1.102/24]):::blueNode
ctrl2-port2-ip([10.42.2.102/24]):::greenNode

card1pif1<-->card1-pif1-ip
card1-pif1-ip<-->vlan421
vlan421<-->ctrl1-port1-ip
ctrl1-port1-ip<-->ctrl1-port1
vlan421<-->ctrl2-port1-ip
ctrl2-port1-ip<-->ctrl2-port1

card2pif2<-->card2-pif2-ip
card2-pif2-ip<-->vlan422
vlan422<-->ctrl1-port2-ip
ctrl1-port2-ip<-->ctrl1-port2
vlan422<-->ctrl2-port2-ip
ctrl2-port2-ip<-->ctrl2-port2

card1pif2<-->card1-pif2-ip
card2pif1<-->card2-pif1-ip
ctrl1 <----> lun1
ctrl2 <----> lun1

linkStyle 0 stroke:#4A90E2,stroke-width:2px;
linkStyle 1 stroke:#4A90E2,stroke-width:2px;
linkStyle 2 stroke:#4A90E2,stroke-width:2px;
linkStyle 3 stroke:#4A90E2,stroke-width:2px;
linkStyle 4 stroke:#4A90E2,stroke-width:2px;
linkStyle 5 stroke:#4A90E2,stroke-width:2px;

linkStyle 6 stroke:#5CB85C,stroke-width:2px;
linkStyle 7 stroke:#5CB85C,stroke-width:2px;
linkStyle 8 stroke:#5CB85C,stroke-width:2px;
linkStyle 9 stroke:#5CB85C,stroke-width:2px;
linkStyle 10 stroke:#5CB85C,stroke-width:2px;
linkStyle 11 stroke:#5CB85C,stroke-width:2px;

linkStyle 12 stroke:#8C8C8C,stroke-width:2px;
linkStyle 13 stroke:#8C8C8C,stroke-width:2px;
```

### Operating procedure

#### 1. Prepare XCP-ng hosts
1. On one of the host servers, make sure that the multipath.conf configuration includes your storage equipment.

This can be found in the file `/etc/multipath.xenserver/multipath.conf`
2. If your equipment is not present, ask the manufacturer for the multipath configuration for GNU/Linux

Add it to the file ```/etc/multipath/conf.d/custom.conf```

:::info
In this case, the configuration will be kept after updates.
:::

3. If necessary, migrate the VMs active on the XCP-ng host in question to another one
4. Reboot the current XCP-ng host on the affected pool
5. Do the same for all servers in the pool

#### 2. Prepare the pool
Make sure that multipathing is enabled on the pool. To do this, go to the advanced configuration of the pool

If this is not the case:
1. Make sure there are **no VMs running** on an iSCSI and/or HBA SR in the pool
2. Activate "Enable multipathing for all hosts"

#### 3. Configure the SR
Proceed with the iSCSI SR configuration as indicated in the [storage documentation](../../storage/#iscsi).

## Fiber Channel (HBA)
### Requirements
* Check that the Fiber Channel cards model(s) is supported via the HCL
* Two different Fiber Channel cards (It can be the same model)
* Two different SAN switches
* Multiple targets per LUN on your storage unit
* Zoning performed

:::warning
Make sure not to mix Fiber Channel speeds.
:::

### Target architecture
#### Target architecture diagram
```mermaid
---
config:
look: handDrawn
theme: default
---
flowchart LR
classDef blueNode fill:#A7C8F0,stroke:#2C6693,color:#000000;
classDef greenNode fill:#A8E6A2,stroke:#3E8E41,color:#000000;
classDef grayNode fill:#D6D6D6,stroke:#8C8C8C,color:#000000;

subgraph server[XCP-ng host]
direction LR
subgraph Card2[Fiber Channel card 2]
direction RL
card2port1[port1]
end
subgraph Card1[Fiber Channel card 1]
direction RL
card1port1[port1]
end
end

subgraph storage[Storage unit]
direction LR
subgraph ctrl2[Controller 2]
direction RL
ctrl2-port1[port1]
ctrl2-port2[port2]
end
subgraph ctrl1[Controller 1]
direction RL
ctrl1-port1[port1]
ctrl1-port2[port2]
end
lun1@{ shape: lin-cyl, label: "LUN" }

end

vlan421[SAN Switch 1]:::blueNode
vlan422[SAN Switch 2]:::greenNode

card1port1<---->vlan421
vlan421<---->ctrl1-port1
vlan421<---->ctrl2-port1

card2port1<---->vlan422
vlan422<---->ctrl1-port2
vlan422<---->ctrl2-port2

ctrl1 <----> lun1
ctrl2 <----> lun1

linkStyle 0 stroke:#4A90E2,stroke-width:2px;
linkStyle 1 stroke:#4A90E2,stroke-width:2px;
linkStyle 2 stroke:#4A90E2,stroke-width:2px;
linkStyle 3 stroke:#5CB85C,stroke-width:2px;
linkStyle 4 stroke:#5CB85C,stroke-width:2px;
linkStyle 5 stroke:#5CB85C,stroke-width:2px;
```
### Operating procedure

#### 1. Prepare XCP-ng hosts
1. On one of the host servers, make sure that the multipath.conf configuration includes your storage equipment.

This can be found in the file `/etc/multipath.xenserver/multipath.conf`
2. If your equipment is not present, ask the manufacturer for the multipath configuration for GNU/Linux

Add it to the file ```/etc/multipath/conf.d/custom.conf```

:::info
In this case, the configuration will be kept after updates.
:::

3. If necessary, migrate the VMs active on the XCP-ng host in question to another one
4. Reboot the current XCP-ng host on the affected pool
5. Do the same for all servers in the pool

#### 2. Prepare the pool
Make sure that multipathing is enabled on the pool. To do this, go to the advanced configuration of the pool

If this is not the case:
1. Make sure there are **no VMs running** on an iSCSI and/or HBA SR in the pool
2. Activate "Enable multipathing for all hosts"

#### 3. Configure the SR
Proceed with the HBA SR configuration as indicated in the [storage documentation](../../storage/#hba).
4 changes: 4 additions & 0 deletions docs/storage/storage.md
Original file line number Diff line number Diff line change
Expand Up @@ -420,12 +420,16 @@ We also suggest to use a folder on the MooseFS cluster as a root path rather tha

Shared, thick-provisioned storage.

For an iSCSI multipath configuration, please follow [these steps](../../storage/multipathing/#iscsi) first.

In Xen Orchestra, go in the "New" menu entry, then Storage, and select iSCSI. Follow instructions from there.

### HBA

Shared, thick-provisioned storage.

For an Fiber Channel multipath configuration, please follow [these steps](../../storage/multipathing/#fiber-channel-hba) first.

You can add a Host Bus Adapter (HBA) storage device with `xe`:

```
Expand Down