Provider: InterServer Issue: Their looking glass shows my prefixes via direct peering. Production traffic routes through CloudCone transit, adding 34ms.
Evidence:
--- looking glass output ---
BGP table entry for 203.0.113.0/24
Paths: (1 available, best #1)
Not advertised to any peer
64512
198.51.100.1 from 198.51.100.1 (192.0.2.1)
Origin IGP, localpref 100, valid, external, best
--- production traceroute ---
traceroute to 203.0.113.1, 30 hops max
1 10.0.0.1 0.4ms
2 192.0.2.17 1.2ms
3 198.51.100.33 35.8ms <-- CloudCone
4 203.0.113.1 36.2ms
Route via CloudCone adds 12ms versus expected direct path. I have opened ticket #OWL-44891. No response in 8 days.
Can anyone else with InterServer peering confirm their looking glass matches reality? I do not want to escalate to their NOC if I am misreading something.
InterServer has a history of this. Their looking glass runs on a "management anycast" node that is not the same as their production edge. I have documented three previous incidents where their route table diverged from production by 6-48 hours.
The capitalized Nouns in their status page are particularly galling. "Optimized Routing" means "we have not deployed the config yet."
[1] See my previous rant in thread c1047 about Time4VPS identical issue.
Wait. You are telling me the looking glass is not even on the same infrastructure as the actual network? That is broken by design! I have seen some things in this industry but this takes the cake! No wonder people dont trust providers anymore! The whole point of a looking glass is to see what the network sees! If it sees something different then it is just a fancy status page with extra steps!
Post a reply
You need an account to reply.
Log in or
register to join the conversation.