[F10-nsp] Multiple ports hanging/dropping on on S50

[David Rodgers]

On that code rev I've experienced similar issues on occasion on certain chassis even with a single switch not stacking .... 


David

On Feb 11, 2010, at 6:18 PM, wwbach at gmail.com wrote:

> We've recently picked up three used force10 SA-01-GE-48T switches.  We're experiencing random issues where is seems like a significant portion of the switch stack drops connections to the devices connected.  This is happening maybe a couple of times a day.    
> 
> Here is some more info:
> 
> I only know enought to be dangerous -- we have a networking consultant working on it who probably is better with cisco gear but has some f10 experience.
> 
> The swithes are stacked using two stacking cables -- 1 to 2 and 2 to 3.  This seems like this is a supported configuration (cascade) according to this web page:
> 
>   http://www.force10networks.com/CSPortal20/TechTips/0057_HowDoIStackSSeriesUnitsTogether.aspx
> 
> but maybe we *need* a cable from switch 3 back to 1 (ring)?
> 
> All the switches are running SFTOS  2.5.2.2 with QOS, Multicast, Stacking and Routing.
> 
> Nothing unusual in the log.  A few interfaces going up and down but not really corresponsing to when we are experiencing issues.
> 
> No errors in "show stack-port diag".
> 
> We took one of the switches out of the mix and things seemed to be better but not perfect.
> 
> When the problem occurs, everything comes back to normal after a couple of minutes.  
> 
> There are several vlans configured.  One vlan where most of the communication is confined to the vlan has not seemed to have any problems.
> 
> When the problem occurs, it seems to affect ports on multiple switches in the stack.
> 
> Spanning tree routing is enabled as is OSPF.
> 
> Another thing worth mentioning is that the we had some issues configurring multicast routing.  We wanted to use sparse mode PIM but couldn't get it working with a cisco router.  We ended up getting a multicast proxy working -- not sure what this really means. 
> 
> Most of the gear attached to the stack is either broadcom of intel chipset.
> 
> Any ideas on how to futher debug this issue or what to look at?  Thanks.  -- Bud
> _______________________________________________
> force10-nsp mailing list
> force10-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/force10-nsp

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/force10-nsp/attachments/20100211/4408d9eb/attachment.html>