conn_error: 500 Can t connect to x.x.x.x:8006 (Connection timed out)\n,

[warning31]

Hello

I have two cluster servers, I close one of them for pmg testing, but the open server gives this error conn_error: 500 Can t connect to 1.1.1.1:8006 (Connection timed out)\n, the open pmg mail traffic continues, but I cannot reach the web interface.

Good Work

 

[dakralex]

Hi!

Could you post the output of pmgcm status? Where do you get the error exactly? What does the network configuration look like?

It seems like that one PMG node (i.e. the 8006 port number is for our PMG instances) tries to connect to another with the IP address 1.1.1.1, but that is a well-known IP address for CloudFlare's DNS server.

 

[warning31]

Hello

I wrote 1.1.1.1 ip as an example. pmgcm status and network setting are below

root@testpmg01:~# pmgcm status
NAME(CID)--------------IPADDRESS----ROLE-STATE---------UPTIME---LOAD----MEM---DISK
testpmg02(2) x.x.x.x node ERROR: 500 Can't connect to x.x.x.x:8006 (Connection timed out) - - -% -%
testpmg01(1) x.x.x.x master S 19:01 0.46 64% 9%
root@testpmg01:~# cat /etc/network/interfaces
auto lo
iface lo inet loopback

auto ens192
iface ens192 inet static
address x.x.x.x/32
gateway x.x.x.x


source /etc/network/interfaces.d/*

 

[dakralex]

address x.x.x.x/32

Can you ping the testpmg02 node from testpmg01? Is this the subnet mask you really want for that host? In my experience, /32 is normally used to refer to a single host without having access to another host in the network.

 

[warning31]

Hello

testpmg02 is closed. I am trying to test, I want it to continue when one of the clusters is closed. There is no problem in sending and receiving mail on the open server, but the web interface does not open. I added a picture.

Good Work

 

[Stoiko Ivanov]

testpmg02(2) x.x.x.x node ERROR: 500 Can't connect to x.x.x.x:8006 (Connection timed out) - - -% -%

This line tells you that testpmg01 cannot connect to testpmg02 on port 8006 - you need to allow this in all firewalls on the way between both nodes (in both directions)

you can verify that it works by using
`nc -v x.x.x.x:8006` on both nodes (replace x.x.x.x by the IP of the other node respectively)

if the issue persists, please check, and share the journal of both nodes:
`journalctl --since '1h ago'` - see `man journalctl`

 

[warning31]

Hello

testpmg02 is already in poweroff status. Mail traffic continues on testpmg01 and mail sending continues, but I receive the errors I shared on the interface. As a result, there may be a failure in the cluster and some servers may be down.

Good Work

 

[Stoiko Ivanov]

testpmg02 is already in poweroff status.

if testpmg will remain offline and you do not plan on activating it again - simply remove it from the cluster on the commandline
`pmgcm delete <nid>` - <nid> is the nodeid that you see in the output of `pmgcm status`

 

[warning31]

Hello

No it will not stay offline it will work again after the problem is solved testpmg02. It will only stay closed when it is faulty and will open when the problem is fixed

Good Work

 

[Stoiko Ivanov]

No it will not stay offline it will work again after the problem is solved testpmg02. It will only stay closed when it is faulty and will open when the problem is fixed

Then leave it in the cluster - as testpmg01 is the master node I don't see a problem (apart from a few glitches in the GUI) - mail-processing should not be affected by an unreachable node in the cluster.

 

[warning31]

Hello

Yes, there is no problem with mail transactions, but the gui does not open, I need to solve why it is happening, I cannot make any settings. I need to connect to the gui while the other server is working

Good Work

 

[Stoiko Ivanov]

Yes, there is no problem with mail transactions, but the gui does not open

This works here - I can login to the running master-node if one of the other nodes is offline - The pop-up informing me that a node is offline shows up - but clicking ok lets me change the configuration.

So I don't understand what exactly is not working for you.

Please share the journal and the logs that might help understanding where the problem is.

 

[warning31]

Hello

There is a timeout on the pages, I am adding the image again.

Good Work

 

[Stoiko Ivanov]

There is a timeout on the pages, I am adding the image again.

no need to add the same picture over and over - please click on either the configuration tab, or the Rule system tab in the left - does this work?
if it does not work - please share the logs as explained above:

if the issue persists, please check, and share the journal of both nodes:
`journalctl --since '1h ago'` - see `man journalctl`

 

[warning31]

Hello

yes there is a problem with the rules, it is loading and it is stuck


root@testpmg01:~# `journalctl --since '1h ago'` - see `man journalctl`
-bash: Dec: command not found