Need clarification regarding nimble's peer persistence behaviour

[RimvydasLT]

Hi,

Just want to get an clarification regarding peer persistence functionality as I'm not able to find anything in the docs. My question would be - if all of the FC ports will go down on any of two nimble arrays in the array group, will this fact initiate ASO? And if the answer is yes, where can I find such things mentioned in the documentation? Thanks.

[Mahesh202]

Hi RimvydasLT

To find detailed information about the ASO feature and its behavior in HPE Nimble Storage documentation, I recommend referring to the below link. This will provide specific details on the ASO functionality, including when and how it is triggered, its impact on sub-volumes, and any additional considerations or best practices for managing Peer Persistence deployments.

https://infosight.hpe.com/InfoSight/media/cms/active/public/HPE_Nimble_Storage_Peer_Persistence_Deployment_Considerations.pdf

The above link explains the scenarios of ASO
refer to page 61 on TESTING PEER PERSISTENCE FUNCTIONALITY

Hope this helps.!!

Regards
Mahesh.

[RimvydasLT]

@Mahesh202 Hi,

Thanks for the link to the article. I reviewed it before posting here. This article says nothing about Nimble's behaviour when it's all FC ports go to down state. I mainly want to know if nimble will initiate ASO when it's all FC ports will go to down state. Thanks.

[Mahesh202]

Hi RimvydasLT

To obtain accurate and up-to-date information regarding the behavior of HPE Nimble Storage in the scenario you described, including the initiation of ASO when all FC ports go down, I recommend contacting HPE Nimble Storage support directly. The support team will have access to the most current information, product updates, and Test case scenarios. they will be able to provide you with the specific details and guidance you need based on your specific configuration and software version.

FC ports down will trigger a controller failover. ASO will not trigger usually in this case and it's expected only if the array is down or not reachable.

Hope this helps.!!

Regards
Mahesh.

[Thomas_Lam_HK]

Please note that Peer Persistence is design for entire array failure scenraio, to ensure data services can served by survied array.  For the array itself we rely on it's own HA mechansim., that's why we always customer has dual FC ports configured on each controller and connect to dual FC SAN fabric for resilient and avoid SPOF.

And back to your original question - Not ASO will be triggered while all FC ports are down (both active and standby controller).

[RimvydasLT]

Hi,

Thank you for the answers. I also contacted support regarding this issue and they also confirmed that ASO will not be initiated on nimble's FC ports down event. But... My colleague swears that he saw ASO on such event. Also support mentioned that if both FC ports on nimble's active controller will go down, active controller will switch to standby controller. Such thing is also not mentioned in documentation. So after such answer I'm not sure, that ASo will not be initialed:( Hope that support will be able to test this scenario in its lab.

[Thomas_Lam_HK]

Support is correct which while all FC ports on the active controller are down and provided that the standby controller has better connectivity, controller failover occurs and those are just standard controller failover mechanism.

However from your original question, if all FC ports are down, including both active and standby controller, with above rule since now the standby controller doesn't has better connectivity, so controller failover will not be triggered and array just stays AS-IS and can’t providing host data access due to all ports down.

And by design ASO won't be triggered on the event of array losing host connectivity, and host connectivity protection is expected to be rely on array HA protection mechanism such as redundant FC ports and dual FC Fabric, etc…

[RimvydasLT]

Strange thing is in all of this story that my colleague swears that he saw ASO initiated on all Nimble's FC ports down event:/ Hope that support will find an ability to test this scenario in their lab to be completely sure.

But in addition to that I'm also not clear with the fact, that when active controller gets all FC ports down, why such event will initiate controller failover to standby controller if nimble does not care about host connectivity. And btw, this thing is also not mentioned in docs, at least I was not able to find nothing mentioned. And if this is true, maybe some logic is still programmed for all nimble's fc ports down event?:)

[Vipin_S]

Automatic failover HPE Nimble Storage automatic failover (AFO) is the failing over of the group management services from the GL to the BGL. This action occurs as a precursor to a Peer Persistence ASO if the GL array becomes unreachable. Only when array becomes unreachable.  After an AFO, the newly promoted GL becomes accessible from the array group’s primary management IP address, and any SSH session to the former BGL that uses the array group’s secondary management IP address is terminated. When the former GL becomes available, it functions as the BGL and becomes accessible through SSH at the secondary management IP address.

So array doesn't failover if the FC ports goes down. The entire array needs to go down for ASO to work. 

Witness will ping only the management network to check the array status.

https://infosight.hpe.com/InfoSight/media/cms/active/public/HPE_Nimble_Storage_Peer_Persistence_Deployment_Considerations.pdf

Please check the complete guide on Peer_Persistence_Deployment_Considerations

[Sunitha_Mod]

Hello @RimvydasLT, 

Let us know if you were able to resolve the issue.

If you have no further query and you are satisfied with the answer then kindly mark the topic as Solved so that it is helpful for all community members.