Start a Conversation

Unsolved

T

30 Posts

1412

November 23rd, 2020 06:00

PowerStore File services

I am confused about below statement about PowerStore File services.

File services only run on the primary appliance. In a multi appliance Cluster, if the primary goes down, the Primary Control Path services will fail over and a new Primary will be selected. File services (SDNAS) will NOT fail over. File services will be down until the old Primary is brought back online. When the old Primary is brought back online, file can be used again. The Primary Control Path services will not failback – they will remain on the new Primary

If primary appliance had malfunction, until it comes back the file system will not be accessed by client? Is that right? or it saying filesystem will be accessible via redundant node or appliance, but management task can't be? Please advise me to understand this better.

Thanks!

43 Posts

December 2nd, 2020 14:00

Hi,

 

Could you share where you saw this?

 

There are many components to this and I think it may be clearer if we separated them out.

 

  1. Each appliance is fully redundant (dual nodes), so even if an entire node fails, the file services automatically failover to the peer node within the same appliance and continue running. Only in cases of a dual failure (both nodes on the appliance going down) would file become inaccessible. Since file services are only available on the primary appliance, they do not failover to a different appliance within the cluster. If both nodes on an appliance go down, the appliance must be recovered in order to restore file connectivity.
  2. The Primary Control Path is related to the management stack only. It has no impact on file or block data access. This has the ability to failover between nodes within an appliance and, if needed, failover between appliances within the cluster. If the Primary Control Paths fails over to another node or appliance, it does not failback (and can failover again if needed).

1 Rookie

 • 

21 Posts

July 26th, 2021 09:00

Hi chenw87,

In the "PowerStore Configuration" training guide (pg 108) it states:

"

NAS deployment is only available on the master appliance in the PowerStore T multi-appliance cluster with storage locality restricted to the hosting appliance.

  • NAS container is installed on both nodes of the PowerStore T master appliance. Nodes are active-active.
  • In a PowerStore T multi-appliance cluster, the NAS container always run on the appliance it was installed on. It does not move with the master node.

"

"Supports only manual failback of NAS server after node recovery." pg 130.

 

I've got to admit, that like tsamuel090590, I too had read that as meaning the whole NAS would be offline. Until reading your reply I hadn't taken into consideration in this scenario that the appliance is two nodes.

 

Simon

(Halfway through reading the PowerStore training guides, aiming for DES-1221)

43 Posts

July 26th, 2021 10:00

Thanks, Simon.

 

Yes, it is definitely important to remember that:

  • Two nodes make up an appliance
  • Both nodes in each appliance are active and can serve IO
  • There can be multiple appliances within a Cluster
  • In a multi-appliance Cluster, File functionality is available on the primary appliance (both nodes)
No Events found!

Top