SAN Zoning Questions

[jon1]

Hi guys,

First time poster, just have a couple of questions if anyone has the time and/or inclination.

1. Someone has previously created zones containing both a single hosts plain pwwn and also the alias'd pwwn. I'm pretty it wont but can someone confirm that no outage will occur to the host if I remove the plain pwwn from the zone but leave the alias?

2. We have old switches left over in fabric manager/DCNM that need to be removed. Please see attached. How can this be removed? Right click>Purge?

3. We have two 9148 switches in a single fabric but we need to change principal switches. Can this be done no distruption?

Thanks in advance!

[Francesco Molino]

Hi 

1. Alias is local to the switch. IAS just cosmetic to have a readable and understandable config instead of talking with pwwn. If zones are not only based on alias and you're sure that pwwn are also in (weird config) you can remove them without disruption. To be sure, with a config we can confirm. 

2. I didn't worked a lot with dcnm and can't answer you. 

3. What do you mean by principal switches? Do you have an actual sketch of the network you want to change and what is the end goal? With that I'll be able to answer you. 

Thanks 

PS: Please don't forget to rate helpful answers

[jon1]

Hi,

Thanks for your reply.

DCNM/fabric manager - not sure of the correct way to remove an old, disconnected switch.

One switch int he fabric is running as the subordinate switch and the other is the principal. I need to swap these around. I'm not sure if this process is disruptive

[Francesco Molino]

What do you mean by subordinate? A switch managing all vsan for example?

Do you have a quick design? you can have multipathing and then it allows that your edge devices could be connected to 2 different "master" switches. In this case not disruptive.

However it depends if everything is Cisco or if there're some 3rd party to take care of the interoperability.

The simplest way is also to prepare your new switch and plan for a migration over a night but this one is disruptive.

Thanks 

[Walter Dey]

see eg.

https://www.cisco.com/c/en/us/td/docs/switches/datacenter/mds9000/sw/5_2/configuration/guides/sysmgnt/nx-os/sysmgmt_cli/domn.pdf

You want all the domain id's to be preserved otherwise, each end device has to go through a flogi, which is disruptive. Therefore make each domain id (one per vsan) static; and then adjust the priority:

    Any new switch can become the principal switch when it joins a stable fabric. During the principal
s  switch selection phase, the switch with the highest priority becomes the principal switch. If two

s  switches have the same configured priority, the switch with the lower WWN becomes the principal

s  switch.

T The required fcdomain restart is disruptive ! 15+ seconds; however there is a fast restart option which           brings this time down to a few msec.

[jon1]

Sorry for the late reply, got sidetracked with another project.

So we have two 9148s, the principal only has 4 ports connected and these devices are not being used/no storage attached. We want to decommission this principal switch and make the current sub the principal but we want to try and do non-disruptively.

Can we simply disable the ISL then decomm the current principal? What will the subordinate do if the ISL is disabled?

I have read about changing fcdomain priority on both switches but this is disruptive from what I have read.

If it helps, here is the output from "show fcdomain" on both switches:

Principal

VSAN 1
The local switch is the Principal Switch.

Local switch run time information:
State: Stable
Local switch WWN: x:2a:6a:79:0a:61
Running fabric name: x:2a:6a:79:0a:61
Running priority: 2
Current domain ID: 0xca(202)

Local switch configuration information:
State: Enabled
FCID persistence: Enabled
Auto-reconfiguration: Disabled
Contiguous-allocation: Disabled
Configured fabric name: x:05:30:00:28:df
Optimize Mode: Enabled (Scale Restart)
Configured priority: 128
Configured domain ID: 0xca(202) (static)

Principal switch run time information:
Running priority: 2

Interface Role RCF-reject
---------------- ------------- ------------
fc1/1 Downstream Disabled
---------------- ------------- -----------

Subordinate

VSAN 1
The local switch is a Subordinated Switch.

Local switch run time information:
State: Stable
Local switch WWN: x:7f:ee:e3:8d:a1
Running fabric name: x:2a:6a:79:0a:61
Running priority: 128
Current domain ID: 0xc8(200)

Local switch configuration information:
State: Enabled
FCID persistence: Enabled
Auto-reconfiguration: Disabled
Contiguous-allocation: Disabled
Configured fabric name: x:05:30:00:28:df
Optimize Mode: Enabled (Scale Restart)
Configured priority: 128
Configured domain ID: 0xc8(200) (preferred)

Principal switch run time information:
Running priority: 2

Interface Role RCF-reject
---------------- ------------- ------------
fc1/10 Upstream Disabled
---------------- ------------- ------------

Any assistance is much appreciated

Cheers

[Walter Dey]

Is it correct, that you use only one VSAN, namely nr. 1 ?

If yes, this is not good and should changed asap. Never use VSAN 1 !

There is a principle switch per VSAN !

I still didn't understand why you want to change the principle switch !

Btw. as long as you have static domain id, and persistent FC-ID, changing the principle switch should not mean a disruption.

[jon1]

Is it correct, that you use only one VSAN, namely nr. 1 ?

If yes, this is not good and should changed asap. Never use VSAN 1 !    - I am aware of this, it was configured like this before my time. This will be done later

I still didn't understand why you want to change the principle switch !    - I said in my post that we wish to decommission the current principal switch

Btw. as long as you have static domain id, and persistent FC-ID, changing the principle switch should not mean a disruption.    - SHOULD not?

Can we simply disable the ISL then decomm the current principal? What will the subordinate do if the ISL is disabled?

Thanks

[Walter Dey]

When you shut the ISL, the principle switch selection process starts automatically on both MDS; you have now 2 FC fabrics, and therefore 2 principle switches ! Therefore take one out of service and your done.

[jon1]

Will the principal switch selection process be disruptive for the current subordinate switch?

Thanks

[Walter Dey]

Because you have static domain id (don't forget, this applies for each VSAN !) and persistent FC-ID, it is non disruptive.

see eg. http://stor.balios.net/Divers/FC/BRKSAN-3708%20-%20Advanced%20SAN%20Troubleshooting.pdf

[jon1]

Thanks for the document

Is there any documentation on the actual principal selection process? Like specifics?

- What will trigger a PSP?

- How long after a PSP event will the process actually trigger?

- How long does the process take?

I cant find anything on specifics like this

[Walter Dey]

see eg. http://www.feeny.org/wiki/index.php/Principal_Switch_Selection

If you change your FC fabric, by adding or removing new MDS (bringing up a ISL link) the principle switch selection will automatically start.

To have static domain id's (per VSAN and switch), you create a stable configuration, independant of time sequence, how switches are powered on (eg. in case of a power failure).

[jon1]

Thanks for the link

But our sub switch (which we need to become the principal) only has a preferred domainID:

Configured domain ID: 0xc8(200) (preferred)

[Walter Dey]

Just change it from preferred to static, and use the same value (200), therefore nothing is changed. It would even work without this change; but make everything static is best practise.