MDS reboot after 497 days uptime

Question

Hey guys,

I just read an important article on the Cisco website about a flaw in some SAN-OS's. Read it on http://www.cisco.com/en/US/ts/fn/631/fn63178.html
The most current SAN-OS EMC recommends is only 3.3 (1c) and this issue is according to Cisco solved in SAN-OS version 3.3(2) and up as well as NX-OS 4.1 (1b) and up.
I have most switches upgraded to 3.3(1c) now, but still this worries me. does anyone know when EMC releases newer versions ?

ConnectrixHelpe · Accepted Answer

Hello,EMC and Cisco are in constant contact in regards to any issues that are found and any field notice they may release. EMC released ETA emc199417 along with Cisco's Field Notice on the issue. The issue is resolved in both 4.1(1b) and 3.3(2) code. NX-OS 4.1(1b) is EMC GA and 3.3(2) is available via RPQ as it is still in testing at EMC. An RPQ is a process that EMC extends support to code before it is generally released. You can contact your local EMC support representative to inquire about an RPQ. Your local EMC support representative can also inform you when EMC plans to release 3.3(2) officially. As it is a future release I am unable to post that information on this board.Also just as some background, the reason for the delay in EMC support of MDS code after Cisco GA's it, is our additional testing. Multiple groups at EMC perform weeks of testing on the code after it is received from Cisco. EMC works to ensure that the code will work with all approved solutions and that serviceability requirements are met. We work with Cisco to find any issues and if necessary get new versions of code that resolve any found issues. This additional testing and documentation takes time and creates the difference in release dates for Cisco versus EMC. Thank you.

dynamox · Answer

i share your concern as well ...as i have four 9222i on 3.3.1. I am not ready to go to NX platform ..so would rather stick with 3.x. I still have 400 days before it reboots but it takes time and planning to scheduled this upgrade as it will bounce the switch

AranH1 · Answer

Using SANTap makes alerts like this difficult as the version of the SAN-OS I can use is limited to what is supported for SANTap, which in RecoverPoint 3.0 is 3.2(2c). Some of my switches are at 387 days

.... time to schedule a reboot.

AranH1 · Answer

CiscoKid,Thanks for chiming in here. I do appreciate the additional testing that EMC performs on the MDS code before releasing it as GA. I think EMC does more interoperability testing across it's entire product line than any other vendor I have worked with and supports a large range of configurations. But the flip side is the time it takes to have releases available for the end users. An understandable delay, but a delay nonetheless.

dynamox · Answer

CiscoKid ..would you please give a glance at 3.3.2 release notes ? I would like to know how i should plan to resolve this issue.

RRR · Answer

I think upgrading the switch can be done online, but the problem is the new code is not GA yet. I suggest a planned reboot or this RFQ or simply wait until GA.

dynamox · Answer

9222i ? When i upgraded from 3.x (whatever came in from the factory) to the latest 3.3.x ..i had to bounce the whole thing, it's only one supervisor so i don't see how can i do it online ?

RRR · Answer

NX-OS = SAN-OS. It's only a name change

ConnectrixHelpe · Answer

It depends on what code level you were at and what the configuration is on the switch if the upgrade will be disruptive or not. Please refer to the EMC release notes for the upgrade / downgrade matrix to see if it is an online or offline upgrade.

Thank you.

RRR · Answer

I think since 3.0 even Cisco can do it online, just like Brocade and McData's. Cisco is just a bit late with this feature. the only thing that's disruptive is the LAN-interface. FC stays online

RRR · Answer

Look at http://www.cisco.com/en/US/docs/switches/datacenter/mds9000/sw/4_1/ssi/upgrade/guide/SSIupgd.html table 1.3

dynamox · Answer

i guess i don't understand Cisco MDS architecture, i thought supervisor was the brain of the switch ..without it it will not function ?

AranH1 · Answer

Really? For minor releases or major releases as well. My experience on the 92xx switches with the single built-in supervisor was that it would reset the supervisor which caused a brief outage.

RRR · Answer

Simple switches like 9124/9134 can be upgraded non disruptively as well. during the upgrade process you'll even be informed about this.

I've done this tons of times and never had FC downtime ! Attached HBA's don't know about the switch being upgraded

Cool hey ?

RRR · Answer

The models you mention are old ones.I know for a fact that upgrades for the 9124, 9134 and 9222i are nondisruptive for FC traffic.

Connectrix

MDS reboot after 497 days uptime

Was this post helpful?