test BC nodes to establish root cause for instability

[ramezsaeed]

• can you please add a second ip/url to check internet connectivity to the monitoring specs? right now we only use google.com, but if website down our nodes are down also.

• zaibon said that the monitoring algo uses google.com to proof that the node has internet connection. last week google.com was not reachable through various isps in europe. so all the nodes connected via those isps showed no internet connection. so thinking to have a fallback url in the monitoring algo.

[zaibon]

@muhamadazmy something to think about, maybe having multiple endpoint for connectivity test is not a bad idea.

[muhamadazmy]

@zaibon yup.

[ramezsaeed]

https://github.com/threefoldtech/0-core/pull/136

[christopherhutton]

do we know this is definitely the root cause of the problem on the BC nodes ? @robvanmieghem @zaibon

[christopherhutton]

until we know the actual root cause then how can we close this issue ? @robvanmieghem @LeeSmet @ramezsaeed pls advise if the root cause is known ?

[sabrinasadik]

• can you please add a second ip/url to check internet connectivity to the monitoring specs? right now we only use google.com, but if website down our nodes are down also.
• zaibon said that the monitoring algo uses google.com to proof that the node has internet connection. last week google.com was not reachable through various isps in europe. so all the nodes connected via those isps showed no internet connection. so thinking to have a fallback url in the monitoring algo.

We thought this was the solution to the bc nodes instability, but even tho this was done, the problem still remains. Nodes are still going offline. Need to find the reason for this so reopening issue.

[sabrinasadik]

Turns out that a lot of the people whose nodes are offline, had a power outtage or unplugged the node to put into another plug, and when they did that the node did not reboot automatically. And some even died because of the sudden power outtage.