thelastman at gmail
Jul 20, 2012, 3:27 PM
Post #3 of 3
(512 views)
Permalink
|
|
Re: recommended procedure for switch maintenance
[In reply to]
|
|
I think the vif favor commands does some sorcery to avoid sending a
reset to clients with established sessions to that vif. Something
like tcp state replication to the other port in the single mode ifgrp.
Something like that...
-Blake
On Fri, Jul 20, 2012 at 2:47 PM, Alon Zeltser <alonz [at] emet> wrote:
> You can use ifgrp favor command to manually switch network activity to one interface while your network guys work on the passive inferface switch
> It will be a graceful transfer rathen then the sudden link break when the switch goes down
>
>
>
>
> Sent from Samsung MobileFletcher Cocquyt <fcocquyt [at] stanford> wrote:
>
> Our networking team needs to do switch maintenance (hardware and firmware)
>
> Each of our 3270 cluster heads is connected to redundant switch pairs - the ifgrp (aka vif) is setup as a single mode:
>
> ifgrp create single na01-vif0 e1a e1b
>
> Current status (abbreviated for readability):
>
> na01-vif0: 1 link, transmit 'none', Ifgrp Type 'single_mode' fail 'default'
> Ifgrp Status Up Addr_set
> up:
> e1b: state up, since 20Jul2012 12:01:50 (00:41:27)
> mediatype: auto-10g_sr-fd-up
> flags: enabled favored
>
> down:
> e1a: state down, since 20Jul2012 12:02:21 (00:40:56)
> mediatype: auto-10g_sr-fd-up
> flags: enabled unfavored
>
> Q: what are the recommended Netapp steps to preserve uptime with the least disruption while the networking folks upgrade each switch in turn?
>
> thanks!
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> Toasters mailing list
> Toasters [at] teaparty
> http://www.teaparty.net/mailman/listinfo/toasters
_______________________________________________
Toasters mailing list
Toasters [at] teaparty
http://www.teaparty.net/mailman/listinfo/toasters
|
