cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Announcements
Field Notice 70545
1326
Views
0
Helpful
5
Replies
mazhar mahadik
Beginner

UCS EMC clarrion boot lun redundancy

Hi Experts,

DESIGN : we have 2 fabric interconnects and 2 san switches conected to EMC clarrion in redundant way.

  fabric interconnect A connects to SAN switch A

fabric interconnect B connects to SAN switch B

we have 2 vHBA for each service profile one with FI-A in VSAN 1010

and other Vhba with FI-B in VSAN 1020

ANd BOOT policy has both of above vHBA going to same target wwwn of boot lun.

We were doing testing for redundancy for SAN connectivity, we faced following issue.

Just a brief, what we had done for testing.

there are two SAN switches , switch A ) and switch B

While all blades are up and running, we switched off SAN switch B first. System was up with no issues.

Connectivity status shows two connections down.

We rebooted all blades and all blades booted up fine. Then we power up SAN switch B.

All connectivity shows ok. All blades working fine.

After 20 min. we power off SAN switch A. All blades were up and running fine.

In the connectivity status we can see two connections down.

We shut down the blades and tried booting. But non of the blades were booting.

We power up SAN switch A and tried booting all blades. All blades came up fine.

From this we can see that we do not get redundancy for boot LUNs.

If switch A is down , we will not be able to reboot the system.

Any solution to this????

Thanks in Advance.

5 REPLIES 5
mikram
Enthusiast

Hi Mazhar,

Are you using Failover mode 0 or Failover mode 1 on Clariion? Is it CX-500/700 storage array?

Basically,

A CLARiiON array is an Active/Passive device and uses a LUN ownership model. In other words,
when a LUN is bound it has a default owner, either SP-A or SP-B. I/O requests traveling to a port SP-
A can only reach LUNs owned by SP-A and I/O requests traveling to a port on SP-B can only reach
LUNs owned SP-B. It is necessary to have different failover methods because in certain situations a
host will need to access a LUN on the non-owning SP.


Are you using Failover mode 1,2,3 on Clariion?

DId you use Navishpere manager to configure LUN?

Hi Mikram,

Thanks for the reply,

basically i m not a EMC guy, i m working on UCS part of th project, so not sure about failover modes on clarrion. San implementation was done by EMC itself, i get back soon on which failover mode is used?

If i am interpreting correctly if failover modes are not correct then it will be active/passice storage, and lun owned by either SPA or B do not have access to each other?

i will check with EMC on failover modes?

I have a similar setup to what you described, but haven't seen that issue happen yet when I reboot either the Cisco FI's or my Brocade Fiber Switches. 

I do have the storage processors on my EMC CX3-40c dual-homed though.  SPA has a connection to each Fiber Switch, as does SPB.  That's how EMC set it up for me anyway.

As far as the failover modes, I found this on the EMC Communties site:

https://community.emc.com/blogs/YahyaAlamlahi/2011/01/08/failover-modes-on-a-clariion-array

Good luck,

Lewis Benton

rewilkin
Beginner

Thanks Lewis, we are waiting to hear back from Makim on his failover. The setup you have works, but when going from fi's to fc switch, we like to see one to one, not bow tie, just easier to troubleshoot, but your setup works fine also. Once Makim checks the failover Mode and ensures it is correct, I believe the issue will be resolved.

Sent from Cisco Technical Support iPhone App

mazhar mahadik
Beginner

Hi folks,

sorry for disappearing for a while,

i have found the solution actually we had messed up initial boot order,

as i explained before i have 2 Vhba and both had same boot targets, for SPA0 & SPA1 ,

san switch A is connected to SPA0 and switch B is connected to SPA1,

so if i switch off switch A then both vhba have same target going to SPA0 which is connected to switch A,

But now we changed target on vhba secondary to SPA1 which is connected to Switch B , so we have switch level redundancy now.

Thanks guys for your help.

Create
Recognize Your Peers
Content for Community-Ad