Fibre Channel (SAN)

Reply
New Contributor
Posts: 2
Registered: ‎04-26-2013

Paths down when principal switch removed

Hi,

In a 3 switch SAN, we removed the principal switch by disabling an ISL, leaving the remaining 2 subordinate switches to agree on which becomes the new principal. During this process we received path down errors on some of our Linux hosts (connected to the 2 remaining switches). I need to understand the principal switch selection process and any fabric re-configuration that occurs as a result, and be able to explain why we got path down errors. Specifically, is there any traffic interruption during this process? Any assistance appreciated.

Thanks.

Contributor
Posts: 66
Registered: ‎03-31-2013

Re: Paths down when principal switch removed

Hello

When there is a change in the topology Fabric Reconfigures, but it is a backend process. No SCN is initiated.

And For SAN Path Error:

Traffic interruption absolutely No.

Did you observe and Link resets on any blades?

New Contributor
Posts: 2
Registered: ‎04-26-2013

Re: Paths down when principal switch removed

Hi Thrinath,

I didn't observe any link resets in the switch log, however the Linux host reported a 5 second outage as follows. The paths on the other fabric were unaffected, so we didn't have an outage. But I still need to understand what happened here.

Apr 24 16:20:23 EIN-1-DB-Prod1 multipathd: mpathb: remaining active paths: 0

Apr 24 16:20:23 EIN-1-DB-Prod1 multipathd: mpathb: Entering recovery mode: max_retries=6

Apr 24 16:20:23 EIN-1-DB-Prod1 multipathd: mpathc: Entering recovery mode: max_retries=6

Apr 24 16:20:28 EIN-1-DB-Prod1 multipathd: mpathb: sde - tur checker reports path is down

Apr 24 16:20:28 EIN-1-DB-Prod1 multipathd: mpathb: sda - tur checker reports path is down

Apr 24 16:20:28 EIN-1-DB-Prod1 multipathd: mpathb: sdc - tur checker reports path is down

Apr 24 16:20:28 EIN-1-DB-Prod1 multipathd: mpathc: sdh - tur checker reports path is down

Apr 24 16:20:28 EIN-1-DB-Prod1 multipathd: mpathc: sdf - tur checker reports path is down

Apr 24 16:20:28 EIN-1-DB-Prod1 multipathd: mpathc: sdb - tur checker reports path is down

Apr 24 16:20:28 EIN-1-DB-Prod1 multipathd: mpathc: sdd - tur checker reports path is down

Apr 24 16:20:28 EIN-1-DB-Prod1 multipathd: mpathb: sdg - tur checker reports path is down

Apr 24 16:20:33 EIN-1-DB-Prod1 multipathd: mpathb: sde - tur checker reports path is up

Apr 24 16:20:33 EIN-1-DB-Prod1 multipathd: 8:64: reinstated

Apr 24 16:20:33 EIN-1-DB-Prod1 multipathd: mpathb: queue_if_no_path enabled

Apr 24 16:20:33 EIN-1-DB-Prod1 multipathd: mpathb: Recovered to normal mode

Apr 24 16:20:33 EIN-1-DB-Prod1 multipathd: mpathb: remaining active paths: 1

Apr 24 16:20:33 EIN-1-DB-Prod1 multipathd: mpathb: sda - tur checker reports path is up

The host is question is Suse 11 service pack 2 on a HP C7000 blade enclosure.

Other blades running Windows VM's also reported path redundancy errors.

So in summary:

Switch A,DID 11 ------ISLx------ Switch B, DID 12 --------ISLy-------Switch C, DID 1 (Principal Switch)

All switches are in different sites, Switch A in Site A, Switch B in Site B, etc.

ISLy shut down for maintenance, isolating Switch C.

Switch A and B decide which is the new Principal switch according to the principal switch selection process.

Host in Site A accesses storage in Site A (no replication) and reports loss of paths when ISLy goes down.

Thanks,

John

Contributor
Posts: 66
Registered: ‎03-31-2013

Re: Paths down when principal switch removed

Hello John

If possible can you pls paste in the Switch logs during this time frame for SwitchA and SwitchB as well

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.

vADC is now Pulse Secure
Download FREE NVMe eBook