[c-nsp] Problems with A99-8X100GE-TR and IOS XR 6.5.3

[Bryan Holloway]

Turns out we had a bad backplane ... ended up replacing the chassis. All 
the LCs were fine.

Thanks to everyone who responded on- and off-list!


On 6/5/23 16:34, Tom Hill wrote:
> On 2023-06-01 21:38, Bryan Holloway via cisco-nsp wrote:
>> https://community.cisco.com/t5/xr-os-and-platforms/problems-with-a99-8x100ge-tr-and-ios-xr-6-5-3/td-p/4013668
>>
>> This chassis has been up and churning packets for several years, so it
>> was presumed to be a possible hardware fail.
>>
>> We acquired another A99-8X100GE-TR and hammered it in, but it won't
>> boot at all. Similar issues -- NPs are unhappy -- won't initialize
>> right.
>>
>> Our fear is a backplane issue, until I stumbled across that post.
>>
>> At the moment, we are in fact running 6.5.3 SP6. FPDs are upgraded, we
>> also have RSP880s, same chassis (9904). Sound familiar?
> 
> I've not seen this specific issue, but I am curious around two things:
> 
>   1. Given the insight from the URL referenced, when did you upgrade the 
> FPDs to the version required by 6.5.3? Presumably after upgrade? How 
> long did it take between that action and the card misbehaving?
> 
>    1a. Is there a difference in HW revision on the 8x100GE card you 
> had/have bought to replace?
> 
> 
>   2. I went to check for obvious things in the release notes for 6.5.x 
> (e.g. a x.y.4 or .5 release that cites a fix for your LC) but it's "so 
> old" a release train that the relnotes have been taken down.
> 
>    2a. Even the FPD values in the 6.7.1 release appear to be newer for 
> the 8x100G card, so I would be remiss for suggesting that you get 
> yourself on a supported release train - as even without hitting 7.x 
> you're still missing out on updates - and probably fixes - for your card.
> 
> Too many years of working with XR caused me to worry a lot less about 
> version upgrades, but to also the scour the release notes very, very 
> carefully before I did any of them.
> 
> Tom