Changed relay IP mid epoch, what are my options

Developers Misc Dev Talk
#1

So I switched my IPs on my relays (wanted to make them elastic) and I made the mistake of pushing up that change… meaning it looks like my BP can’t recognize my relays until 2 more epochs. Is there anything I can do to get around this? And in the even that I ever had to change my IPs in the future, how would I get around this?

#2

Hi,

Changing the IP will not affect the static connection configuration

So it will should work as long the new IPs are modified on BP topology file and also on FW

1 Like
#3

Okay so you are saying I should be good then?
also any idea why telnet from my relay to bp works when I use the bp private ip, but connection times out when I use the elastic public ip? When I run gLive everything is looking good, but when I manually check connections between relay and bp it doesn’t appear to be happening. (I am using aws security groups to do this)

#4

Confirmed, I tried the reachability analyzer on aws and it only works for the private ip and not the elastic ip between relay and producer

1 Like
#5

The nodes are connected or not?

#6

Perhaps u will need a PF on ur local router? The elastic IP it’s using the public internet and if the port is not opened on FW then is normal to not work

1 Like
#7

Why PF on local router? I can ssh in to the elastic ip no problem, its just the communication between the relay and producer. The security groups seem good because I am gettin proper behavior between relay and producer on private ip, it’s just something with elastic. I don’t think I have an FW on the OS as my iptables look empty

1 Like
#8

How do I tell? not sure what you are asking

#9

Alternatively, you can specify the Elastic IP address in a DNS record for your domain, so that your domain points to your instance

Did u tried with DNS?

#10

So how is my producer running if it isn’t communicating with my relays?

#11

I haven’t tried dns, would I just use the aws provided dns to test?

#12

You should try for test if it’s free but let me understand … ur nodes are on aws and they are on the same network?

#13

Same aws account, same region, different instances.

security group 1: allows ssh connection
security group 2: allows all tcp to relay port
security group 3: custom tcp to relay port from security group 1

relays: security group 1 and 2
producer: security group 1 and 3

so weird to me that these security groups are working well for private ip but not the public elastic

#14

Did u contacted the cloud support team? Perhaps additional setup is required? Each server has a different elastic IP assigned right?

#15

I haven’t yet, was really hoping it was something dumb I did. Yes different elastic IP for each server, even tried a new one on the producer and changing the port in the security group

#16

Ok, then… check the ssh port for eac port from here

Do u see it opened?

#17

So that is saying they are closed

#18

however the port I have designated to be open on my relays does show open

#19

So, the cnode port is showing opened on both relays right? They have IN peers right?

ur topology should looks like:

  • BP to relays and relays to BP use private network
  • Relays to other public nodes - public nodes to relays use public IPs (elastic IP)
#20

relays both show open, still very new to this so not sure if they have IN peers. All topologies are using the public elastic IP

BP topology:
{
“Producers”: [
{
“addr”: RELAY2_IP,
“port”: RELAY1_PORT,
“valency”: 1
},
{
“addr”: RELAY2_IP,
“port”: RELAY2_PORT (same as relay 1),
“valency”: 1
}
]
}
relay topologies:
{
“Producers”: [
{
“addr”: PRODUCER_IP,
“port”: PRODUCER_PORT,
“valency”: 1
},
{
“addr”: “relays-new.cardano-mainnet.iohk.io”,
“port”: 3001,
“valency”: 2
}
]
}