Primary Storage

 View Only

 Flashsystem, PBHA: "IP connectivity is required between systems for management. "

Ralf Lichtenstein's profile image
Ralf Lichtenstein posted Wed December 04, 2024 06:35 AM


Hi community,

please allow for the following questions.

The documenation on PBHA (https://www.ibm.com/docs/en/sanvolumecontroller/8.7.x?topic=availability-planning-high) states:

"IP connectivity is required between systems for management. The following TCP ports are required to be accessible on the system management IP address:

    Port 7443 for REST API access between systems.
    Replication setup using the management GUI requires access to both systems from the host where the web browser is running. Ensure that any firewalls between the web browser and storage systems allow inbound traffic to port 7443 and 443 on the system management IP address."
    
    
Questions:

- Is tcp/ip connectivity between PBHA member systems required _all the time_ after creating the partnership?

- What exactly will break if the tcp/ip connectivity between PBHA member systems breaks for some reason?

Regards

Ralf

Nezih Boyacioglu's profile image
Nezih Boyacioglu

Hi Ralf,

It's the same as FC connectivity: if you lose the connection between sites, ip quorum checks which site is the preferred site and takes the non-preferred site offline to prevent potential data inconsistencies. You can check the IBM Redbooks for PBHA https://www.redbooks.ibm.com/abstracts/sg248569.html

Philip Clark's profile image
Philip Clark

Losing connectivity does not break HA, but it will affect management activity.

If there is an interruption in IP connectivity, HA will continue running, but you will not be able to make any changes, including provisioning new volumes, etc.

Ralf Lichtenstein's profile image
Ralf Lichtenstein

Nezih, Philip - thank you for taking the time to answer my question.  But: who is right? Can you please clarify?

Uncertain regards

Ralf 

Philip Clark's profile image
Philip Clark

The comment about IP connectivity working the same as FC connectivity, and failing over to the preferred site, is referring to using Ethernet partnership for HA.

That is not true for the REST API management connection.

Ralf Lichtenstein's profile image
Ralf Lichtenstein

Hi Philip,

thank you.  Your answer is exactly what i _thaught_  (or maybe _hoped_) it would be. Althought i could not find any documented information on this.

Please allow for one (hopefully) last question: where exactly did you get this information from? 

Regards

Ralf

Nezih Boyacioglu's profile image
Nezih Boyacioglu

Oh, it's about REST API connection, sorry, misunderstood and thought it was about Ethernet based ha setup.

Philip Clark's profile image
Philip Clark

I asked development, but you can find it documented here: https://www.ibm.com/docs/en/flashsystem-9x00/8.7.x?topic=replication-high-availability-disaster-recovery-3-site

This is a 3-site page, but there is a row that covers: Loss of management IP connectivity only between the HA systems

Ralf Lichtenstein's profile image
Ralf Lichtenstein

Hi Philip,

thank you very much. I appriciate that. 

Regards 

Ralf