Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

2332

September 28th, 2009 12:00

MDS 9222i Code Upgrade

Hi All. It has been a long time since my last post...!

I am in the process of upgrading my MDS 9222i switches from 3.3(1c) to probably NX 413a (prob a "period" in there somewhere but not sure where :-).

The existing code has been very good with the exception of the well known 497 day reload, which is the only reason why I am upgrading.

1. Who has done an upgrade similar to mine (HW, existing and target code, etc)?

2. Any issues with a NDU upgrade? My gut tells me to reload (warm and fuzzy feeling to see the uptime reset to zero) after any type of upgrade but if it is truly safe then maybe I will not.


Thanks guys!

2 Intern

 • 

20.4K Posts

September 29th, 2009 11:00

You mean there is a place to check compatibility? ;-)
I don't check, that is why I am posting.


sure, just add your components (swithces/arrays/host/powerpath) to E-Lab navigator and it will tell you where you need to be code/firmware/version wise.

Is that a joke, did not know Qlogic plays in that
space.


not a huge player compared to Brocade/Cisco ..but they do have nice switches.

2 Intern

 • 

20.4K Posts

September 29th, 2009 12:00

I was just kidding about e-lab.

So, which switch do you like best? Lets say it
together now, C-I-S-C-O...

Or should I ask you after 0ct 17?


i've had good experiences with both vendors, Cisco does have a lot of bells and whistles though.

2 Intern

 • 

5.7K Posts

September 29th, 2009 23:00

True reboot as in: ethernet down as well as FC down ! Just like you switched of the entire switch and put it back on. I meant downtime on that switch !

For the rest all upgrades were true NDU !!! Very good !

2 Intern

 • 

5.7K Posts

September 29th, 2009 23:00

I say B-R-O-C-A-D-E for president !!!! Cisco has way too much bells and whistles. In the FCoE field however Cisco seems to be a major player, but in FC I think Brocade is the winner.

September 30th, 2009 09:00

Hey RRR.

Did the reboot happen right away?

October 1st, 2009 09:00

Must be user error ;-) J/K

I assume the ones that did not reload, the counter is still incrementing and past 497 at this point? I think you talked me into it, NDU all the way.

It's kinda like hotswap memory chips on servers, who would really want to do that...!

Thx

2 Intern

 • 

5.7K Posts

October 2nd, 2009 02:00

Good question ! I must say that I never stayed around at the customer's site to see if the counter actually now goes over 497 ;) I think my boss would kill me. I have other thnigs to do than watch counters ;)

October 2nd, 2009 07:00

I see, so you are a CE or a consultant of sorts? Well, if they didn't call you I'm sure things are good.

Thx

2 Intern

 • 

5.7K Posts

October 5th, 2009 00:00

Consultant, working for an EMC ASN partner.

October 12th, 2009 14:00

Well, I had a very fun day last week upgrading my switches to NX code. Let me tell you something, it was not NDU (I did have to reload).

Turns out the install script was unable to install erroring out with a 0 or 1 return code. After working with the local CE and a level 1 guy on the phone, they were lost! The guy on the phone was typing the same command over and over believing it will run at some point, not.

The call was escalated and got a more SR level guy. He seen this issue before and told us that the current code is hooked to a SSI module that is not even in my system, see below for a "show mod":

Mod Application Image Description Application Image Version
-------- ----------------------------- -------------------------
1 SSI linecard image (Packaged in SAN-OS) 3.3(1c)

I was told reloading the switch will clear this issue, which it did. We then proceeded with the install.

Second switch did the same thing BUT the module was still there, what joy. After about an hr I suggested to try the install anyways and it actually worked! If you did a show mod after the upgrade, the SSI mod was no longer there. Come on, seriously...!

The Sr level tech stated he has seen this many time before. I wish I had this information to begin with. Although the reload did not cause any issues since all but one host is dual connected, I was not ready for this.

Looking at my prod switches, I have the same SSI modules, which means I will prepare for a reload as well.

What is interesting is that these set of switches are in different locations and appear to have the same issue. This tells me it's more of a bug, and not a previous upgrade issue from a CE (loading code for something I don't have).

Hey dynomox, put in reload for your change control :-)

2 Intern

 • 

5.7K Posts

October 13th, 2009 03:00

Did you read the release notes ? For some linecards the upgrade isn't NDU indeed !!

2 Intern

 • 

5.7K Posts

October 13th, 2009 05:00

Furthermore: if you upgrade from a somewhat older SAN-OS, you might have to upgrade to an intermediate SAN-OS before upgrading to NX-OS 4.1(3a) or something.
The release notes provide further information on specific configurations such as modular switches with older linecards.

2.2K Posts

October 13th, 2009 12:00

Did you read the release notes ? For some linecards
the upgrade isn't NDU indeed !!


Based on what he wrote though, the problem is that he doesn't have that linecard in the switch. So the upgrade the current code he was on was erroneously reporting that there was a SSM lincard in the switch.

October 13th, 2009 17:00

That's correct, I do not have this linecard.

The engineer is going to write a kb article on it. Too bad he did not write it sooner since he has seen this problem before.

2 Intern

 • 

5.7K Posts

October 14th, 2009 00:00

Oops, where was I, what was I thinking..... I was fast browsing through the message and only noticed "...line card...". Excuse me. I must have been asleep or something.
No Events found!

Top