Author Topic: Systems lose heart beats - IP trunks fail  (Read 2477 times)

Offline VinceWhirlwind

  • Hero Member
  • *****
  • Posts: 899
  • Country: au
  • Karma: +31/-0
    • View Profile
Re: Systems lose heart beats - IP trunks fail
« Reply #15 on: February 26, 2018, 07:14:27 PM »
What do you mean by 'aggregated links"?
That would be where a "link" between two devices is formed using 2 or more physical links, configured to act as a single virtual link using LACP usually. ("Trunks", "Etherchannel", "802.3ad", various names for it depending on vendor).
I have (rarely) had comms affected by a bug in link aggregation whereby after some random amount of time the aggregated link stops passing *some* traffic. I've had this on Nortel as well as VMWare.
To rule out such a bug, disable all physical links in the aggregated link except for one and monitor for a period
 
We only lose connection between these two controllers.  These two servers are the must busy in the cluster.  The ISS server has ~200 users and the MCD controller has all the trunks - both SIP and T1.

Doesn't sound exceptionally busy, however that's all a matter of perspective from the point of view of the switch - that switch could have some kind of per-protocol rate-limiting configured on its in-path interfaces and it is blocking traffic after it reaches the configured limit. Bit of a stretch really, but you never know, those Nexus switches can be a bit odd, I haven't used them for years.


Online ralph

  • Mitel Forums Admin
  • Hero Member
  • *****
  • Posts: 5767
  • Country: us
  • Karma: +469/-0
  • Published Author: http://amzn.to/2dcYSY5
    • View Profile
Re: Systems lose heart beats - IP trunks fail
« Reply #16 on: February 27, 2018, 07:41:11 AM »
Thanks for the replies.
I took another look at the MCD controller yesterday.  I can see that it lost connections to all the other controllers on Saturday.  No work was being done on the network.
They just happen to be scheduled for an upgrade in just a little over 2 weeks.
The hardware will be replaced at that time.
We're going to leave a ping against that controller until then.

Ralph


 

Sitemap 1 2 3 4 5 6 7 8 9 10