[c-nsp] Nexus 5k ISSU

[quinn snyder]

On n5k — when designed appropriately — I’ve not had an issue with ISSU.  The process is actually pretty stable.  This means adhering to design guides for things like bridge-assurance, STP, etc.

N7K was a rockier experience (when I was in the field).  5.0 -> 5.1 -> 5.2 transitions were rough — due to scheduler rebuilds with each release.  Many things broke.
I can’t think of any of my customers today running vPC on n7k — as most of them have moved to either ACI or VXLAN-EVPN — so I don’t have any anecdotes from the field.

q.

—
Quinn Snyder | snyderq at gmail.com 

-= Sent via iPad.  Please excuse grammar, spelling, and brevity =-

> On Nov 1, 2019, at 17:39, Bradley Ordner <bradinusa at hotmail.com> wrote:
> 
> I have done this on the 7K and I don’t trust it anymore. I had OSPF adjacencies go down when the supervisor failed over. 
> 
> We plan for outage now, we only have one per DC :( and do it manually. 
> 
> Even running the ISSU commands to see If the device was ready failed sometimes. 
> 
> What I would suggest, which we tried as well to no effect is to reboot the supervisors or what ever the 5k brains are called one by one before trying ISSU. That way it’s fresh.
> 
> Brad Ordner
> 
> 
> 
> Sent from my iPhone
> 
>> On 2 Nov 2019, at 9:19 am, harbor235 <harbor235 at gmail.com> wrote:
>> 
>> Hi everyone,
>> 
>> What are your experiences with Nexus5K ISSU and VPCs.  Do you see service
>> interruptions? ISSU is never quite ISSU. During role changes and/or VPCs
>> reforming I see short duration losses. Is this standard?
>> 
>> 
>> Mike
>> _______________________________________________
>> cisco-nsp mailing list  cisco-nsp at puck.nether.net
>> https://puck.nether.net/mailman/listinfo/cisco-nsp
>> archive at http://puck.nether.net/pipermail/cisco-nsp/
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/