1827276 Members
2006 Online
109717 Solutions
New Discussion

MSA1000

 
trungvt
Occasional Advisor

MSA1000

Hi all,
Pls teach me,
- Why HA in MSA1000 use SAN Switch 2/8 has not ISL?
- How to HA in MSA1000 operate?
Thanks so much!

8 REPLIES 8
Steven Clementi
Honored Contributor

Re: MSA1000

The MSA1000 does not support ISL connected SAN switches. Your High Availability comes from having multiple paths to your storage, 1 path per controller. 1 HBA per switch, 2 HBA per server = 2 paths to your storage.

In other words, the MSA does not support a single HBA "seeing" both controllers.

The MSA controllers are not Active/Active. One can operate at a given point in time. While one is active, the second is in standby mode.

I am sure others will want to add their experiences and interpretation of what you are asking.

Hope this helps somewhat.

Steven
Steven Clementi
HP Master ASE, Storage, Servers, and Clustering
MCSE (NT 4.0, W2K, W2K3)
VCP (ESX2, Vi3, vSphere4, vSphere5, vSphere 6.x)
RHCE
NPP3 (Nutanix Platform Professional)
trungvt
Occasional Advisor

Re: MSA1000

Steven: Thank you for quick your support!
- But I would know more, when controller is in active mode is down, how to controller is in standby mode is actived? what is do it?
- In SAN Data Availability topologies, each level has ISL. when ISL will be used?

Hope you understand my opinion :)

TrungVT
Uwe Zessin
Honored Contributor

Re: MSA1000

Standby controller either becomes active on its own or is told by a host to do a failover.

Can you show us the document that requires each "level" (what's a level?) to have an ISL? A true high availibility configuration uses two different fabrics (at minimum two switches not connected with each other).
.
Steven Clementi
Honored Contributor

Re: MSA1000

- But I would know more, when controller is in active mode is down, how to controller is in standby mode is actived? what is do it?

In the case of the Windows Operating System, which I am most familiar with... Secure Path OR the basic MPIO (Multipath I/O) DSM automatically detect path failures and force the controller roles to swap. There is no supported manual way to get the controllers to swap roles except for maybe disabling a fibre connection or pulling out the fibre.


- In SAN Data Availability topologies, each level has ISL. when ISL will be used?

An ISL is used to connect 2 switches together to form 1 distinct fabric. The MSA requires 2 seperate fabrics to be in a supported configuration (unless that has chnaged recently).

ISL's would be used to:

1. Connect different sites. Assume site 1 has 2 fibre switches and site 2 has 2. You have nodes of a cluster in site 1 and site 2 with storage in site 1. In order for the nodes in site 2 to see the storage in site 1, an ISL would have to be inplace so that the 2 switches can talk to each other and allow traffic to pass back and forth.

2. Connect more switches to increase port count. Assume you have 1 switch with 8 ports and you have 9 hosts and 1 msa1000. You need a few extra ports to connect all the hosts to the msa.

3. Merge different Stand Alone SAN Infrastructures into 1 consolidasted SAN

Hope all this is helpful.

Steven
Steven Clementi
HP Master ASE, Storage, Servers, and Clustering
MCSE (NT 4.0, W2K, W2K3)
VCP (ESX2, Vi3, vSphere4, vSphere5, vSphere 6.x)
RHCE
NPP3 (Nutanix Platform Professional)
trungvt
Occasional Advisor

Re: MSA1000

Steven: Thank you for clear support!
Uwe: You can refer to the following URL to know HA level design and tks for answer:
http://h18006.www1.hp.com/products/storageworks/san/documentation.html

TrungVT
Uwe Zessin
Honored Contributor

Re: MSA1000

I am familiar with this page and I have a good collection of these manuals, but I really don't have time to go through all of them to challange the claim.
I'd rather like to see a document number and page.
.
trungvt
Occasional Advisor

Re: MSA1000

Uwe: At page 59 in SAN design reference guide.
Uwe Zessin
Honored Contributor

Re: MSA1000

Document: AA-RMPNY-TE

Ah, that's better and your question about 'levels' makes sense to me, now.

In level 1, ISLs are simply for connectivity. In level 2, you see additional (redundant) ISLs for resiliency. Level 3 adds redundant paths for the devices.

The problem in Level 3 is that you can still cause problems to the entire fabric, e.g. by making a zoning error.

Level 4 uses at least two completely independent fabrics. An error in one fabric has no effect on the other one. Devices should continue to work if they are connected to both fabrics and have multipath capability.

Traditionally, level 4 was called a high-availability configuration. Looking at table-8, I wonder if it is now called the "highest-availability configuration" ;-)


For a two-controller MSA1000 you need level 3 with proper zoning or, better: level 4.
.