I have a link that for reasons yet to be understood goes into hiding (disconnects) and does not reconnect until ~45 minutes later a watchdog decides to reboot to get access to the POP …
Is there a setting somewhere that could cut down the profound delay ?
There is no such setting. This is a reboot of last resort. If a low timeout is used and for some reason POP is not functioning well, the whole network will reboot and debugging will become a lot difficult. For that reason we kept the timeout at 45 mins and did not give a configurable option.
However I hear you. I will check with the team if we can improve something.
Hi Glenn,
Not sure why this wasn’t configurable, I’ve asked the team to add an advanced configuration control to allow the user to change this.
We are just finalizing the changes to the next release, we’re likely to have a Beta 4, I’ll see if we can get this in there.
So why is this going off, why are you losing the connection, are you a single or multi POP on this network. Would you like us to take a deeper look?
I would be delighted to have your help with this.
Single POP two CNs stay online happy as clams while two DNs lose connection and eventually trigger the watchdog, seems to happen about every three days or so, no consistency as to time of day.
Does Karthik have access to your network or cnMaestro?
Either that or could you raise a ticket and add the field diags from the E2E and the node that reboots.
As you may expect this is only meant to happen if the link goes down and can’t recover on it’s own, I would hope the link never goes down let allow stays down.
I have had no contact with Karthik but that sounds handy
OK please raise a ticket, let me know the number and I’ll get someone to help you. Access to your units would help, either through cnMaestro or VPN. I’m sure we’ll sort this out quickly, the fact it’s going down so often we need to look at the topology.
I had raised a ticket 35053, and sent you email on this.
Just had another blackout event got notified and checked on it within a few minutes, the links showed as up but both DNs downstream were offline, sent a disassociate command to each and BAM up and running again.
Had 8 days of runtime, noticed in the notifications a declaration of a couple links being ‘UP’ at 1:34 AM four days ago, nothing else going on, no watchdog or the like …
