Add cluster sizing quorum warning and link #2583
Conversation
|
@mbrat2005 : Thanks for your contribution! The author(s) have been notified to review your proposed change. |
|
Learn Build status updates of commit c34f88c: ✅ Validation status: passed
For more details, please refer to the build report. For any questions, please:
|
|
Learn Build status updates of commit bf4174f: ✅ Validation status: passed
For more details, please refer to the build report. For any questions, please:
|
#label:"aq-pr-triaged" |
prmerger-automator
bot
added
the
aq-pr-triaged
| @@ -72,6 +72,9 @@ Here are the Azure requirements for your Azure Stack HCI cluster: | |||
|
|
|||
| A standard Azure Stack HCI cluster requires a minimum of one server and a maximum of 16 servers. | |||
|
|
|||
| > [!IMPORTANT] | |||
There was a problem hiding this comment.
This would be better as a bullet rather than an Important reader element. I'll make a recommendation below.
| @@ -72,6 +72,9 @@ Here are the Azure requirements for your Azure Stack HCI cluster: | |||
|
|
|||
| A standard Azure Stack HCI cluster requires a minimum of one server and a maximum of 16 servers. | |||
|
|
|||
| > [!IMPORTANT] | |||
There was a problem hiding this comment.
| > [!IMPORTANT] |
| @@ -72,6 +72,9 @@ Here are the Azure requirements for your Azure Stack HCI cluster: | |||
|
|
|||
| A standard Azure Stack HCI cluster requires a minimum of one server and a maximum of 16 servers. | |||
|
|
|||
| > [!IMPORTANT] | |||
| > When designing the number of nodes in your cluster, make sure to consider how both [Windows Failover Cluster and Storage Spaces Direct quorum](quorum.md) will impact your cluster's availability. | |||
There was a problem hiding this comment.
| > When designing the number of nodes in your cluster, make sure to consider how both [Windows Failover Cluster and Storage Spaces Direct quorum](quorum.md) will impact your cluster's availability. |
| @@ -72,6 +72,9 @@ Here are the Azure requirements for your Azure Stack HCI cluster: | |||
|
|
|||
| A standard Azure Stack HCI cluster requires a minimum of one server and a maximum of 16 servers. | |||
There was a problem hiding this comment.
| A standard Azure Stack HCI cluster requires a minimum of one server and a maximum of 16 servers. |
| @@ -72,6 +72,9 @@ Here are the Azure requirements for your Azure Stack HCI cluster: | |||
|
|
|||
| A standard Azure Stack HCI cluster requires a minimum of one server and a maximum of 16 servers. | |||
|
|
|||
| > [!IMPORTANT] | |||
| > When designing the number of nodes in your cluster, make sure to consider how both [Windows Failover Cluster and Storage Spaces Direct quorum](quorum.md) will impact your cluster's availability. | |||
|
|
|||
| Keep the following in mind for various types of Azure Stack HCI deployments: | |||
|
|
|||
There was a problem hiding this comment.
| - An Azure Stack HCI system requires a minimum of one server and a maximum of 16 servers. If you're using 2-4 servers, we recommend using a [witness](quorum.md) to help determine which server is most up to date if there is contention. |
There was a problem hiding this comment.
Thanks @JasonGerend!
The trigger for opening this PR was that a customer did not understand that Windows Failover Cluster and Storage Spaces Direct manage quorum separately. Whereas a Windows Failover Cluster allows you to configure an external witness resource for maintaining quorum, S2D does not. I worry that changing the wording as above will result in customers configuring WFC quorum but not considering how a node failure could impact S2D availability. I couldn't think of how to spell this out concisely and fell back on linking to the quorum article...
Another way to address this could be a page dedicated to cluster sizing, multiple clusters, and related considerations-- which customers also seem to be asking for
There was a problem hiding this comment.
Good feedback on that table - I agree it's a problem and that we're not yet clear enough on pool quorum. I'm asking some of the PMs for more info and will keep you posted! :)
There was a problem hiding this comment.
@JasonGerend Did you have any luck getting more information from the PG?
There was a problem hiding this comment.
@JasonGerend Have there been any updates on this conversation from the PG?
Add an [important] message and link to the Server Requirements section of the system requirements article to raise visibility in how node count and quorum interact to impact availability. For example, a 2 node cluster will be brought offline by a node failure 50% of the time.