Print

Print


Frank,

 

                I have been gradually implementing the unknown unicast rate limiting (starting with my own extension) to make sure it does not cause any unexpected problems.  I had not applied it to the route servers yet, so I can understand why you saw another blip when a member with a significant amount of traffic had their port go down unexpectedly today.

 

Jeremy

 

From: MICE Discuss [mailto:[log in to unmask]] On Behalf Of Frank Bulk
Sent: Wednesday, April 21, 2021 3:05 PM
To: [log in to unmask]
Subject: Re: [MICE-DISCUSS] Zayo Belle Plaine - April 15th

 

We saw another BFD event at 2:04 pm today – was there a disturbance in the force? 😉

 

And to follow up on my previous question, is Arista falling bit short in our situation, by not supporting a receive interval of 10 msec?

 

Frank

 

From: MICE Discuss <[log in to unmask]>
Sent: Friday, April 16, 2021 4:43 PM
To: [log in to unmask]
Subject: [External GPC] Re: [MICE-DISCUSS] Zayo Belle Plaine - April 15th

 

In this case it was BFD that kicked in – I have BFD only set up for the two route reflectors.  Other BGP sessions stayed up.

 

I also got this alert from our Arista 7504:

 

Apr 15 17:37:34 207.32.15.14 (sxct-north-2.sxcy.fbnt.netins.net) Bfd: %BFD-5-INCOMPATIBLE_RX_INTERVAL: Received a BFD rx interval from peer (vrf:default, ip:206.108.255.2 (AS53679-MRS-2.micemn.net), intf:Port-Channel2, srcIp:0.0.0.0, type:normal) of 10 milliseconds, outside of the supported range of 50-60000 milliseconds. (message repeated 3 times in 1.18783e+06 secs)

 

Is Arista falling bit short in this situation, by not supporting 10 msec?

 

Frank  

 

From: MICE Discuss <[log in to unmask]> On Behalf Of Jeremy Lumby
Sent: Thursday, April 15, 2021 10:52 PM
To: [log in to unmask]
Subject: Re: [MICE-DISCUSS] Zayo Belle Plaine - April 15th

 

Frank,

 

Ideally it should not have, however what tends to happen is that when large ports go down the MAC addresses leave the table, and therefore the traffic gets flooded to all ports until BGP times out on the networks sending the traffic.  Since the route servers only have 1G ports, the 10G of traffic that typically is flowing to Belle Plaine probably saturated the route server ports for a short amount of time.  If you have short timeouts, or use something more sensitive like BFD I could see how your sessions dropped from the saturation.

 

Jeremy

 

From: MICE Discuss [mailto:[log in to unmask]] On Behalf Of Frank Bulk
Sent: Thursday, April 15, 2021 9:40 PM
To: [log in to unmask]
Subject: Re: [MICE-DISCUSS] Zayo Belle Plaine - April 15th

 

We (AS18883) saw our BGP sessions with the MICE router reflectors bounce at that time – should that have happened?

 

Frank

 

From: MICE Discuss <[log in to unmask]> On Behalf Of [log in to unmask]
Sent: Thursday, April 15, 2021 6:05 PM
To: [log in to unmask]
Subject: [MICE-DISCUSS] Zayo Belle Plaine - April 15th

 

Looks like the Zayo Belle Plaine remote went down at 5:37pm, I believe Lee @ SW MN Broadband is opening a ticket now.

 

From the MICE side we have light but no frames, and looking at other unrelated systems it appears the cut is South of Hwy 212 someplace.

 

Best of luck everyone,

Anthony

 

 


To unsubscribe from the MICE-DISCUSS list, click the following link:
http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1

 


To unsubscribe from the MICE-DISCUSS list, click the following link:
http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1

 


To unsubscribe from the MICE-DISCUSS list, click the following link:
http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1

 


To unsubscribe from the MICE-DISCUSS list, click the following link:
http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1

 


To unsubscribe from the MICE-DISCUSS list, click the following link:
http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1



To unsubscribe from the MICE-DISCUSS list, click the following link:
http://lists.iphouse.net/cgi-bin/wa?SUBED1=MICE-DISCUSS&A=1