Resolved Browser- and SSH-session - connection-aborts

[Goodfred]

Hi! I hope you can help me!

I have a fresh installation of Ubuntu 20.4 with a fresh installation of Plesk.

I used the migration manager to copy data from an old Plesk installation to the new server.

I dont know anymore if the problem was there before the migration or it came only after the migration. But I think it was before, too.

The problem:
Sometimes the server is not accessable from the browser. It could happen that Im surfing in the Plesk panel and the browser doesnt answer anymore. To solve this I have to wait (some seconds to minutes) and then I can continue the browsing from there where I was. (session is not killed)
The same problem I have with SSH on the server. When I connect with Putty it could be that the connection aborts. Then I have to reconnect. When I cant access with SSH I cant access with the browser at the same time. (and not on SSH when the browser doesnt work)
The server is not rebooting. (I saw that on the remote console with ESXi. Nothing appears on the screen on the ESXi remote console - everything totaly normal)
And while the browser/SSH is not accessable ping is working! Without lags etc.

I hope you have tipps for me what I can do!

Im very thankful and I wish a good week!

Goodfred

 

[Nik G]

hello @Goodfred ,

I would suggest you to check fail2ban settings: it looks like it is configured in a very strong way and blocks users due to auth fail in web session, for example.
As for SSH - by default fail2ban adds the whole IP to the banned list in case reoccurring errors.

 

[Goodfred]

Hi!

It seems like it is not a fail2ban issue.
I had a connection problem 2min ago and in the log file I cant find a ban for that time.
Do you have more ideas? What can I do?

Thank you for answer!
Goodfred

P.S.: Ping is working while I cant connect
P.P.S.: The error message of the browser is "ERR_CONNECTION_REFUSED" - Does this help?

 

[Nik G]

@Goodfred,
probably it make sense to install an extension called "Advanced Monitoring".
It would help to understand what level of the resources usage all the time when you experience connectivity problems.
The idea is that server may be overloaded during these periods (and I'm talking not only about high CPU usage.
It may be disk I/O or RAM utilization).

Another one idea is a problem with network connectivity in general (as on the server side, as on your side).
As a first troubleshooting step I'd recommend you to connect to the server using another provider (for example switch to mobile internet if it is possible) when you will face such problems again.


If you would not be able to find the reasons you always may contact Plesk Technical Support.

 

[Goodfred]

Advanced Monitoring is showing nothing at the time when the connection aborts.

I think the network connectivity is ok because I can ping whil its not accessable.
But I can try it with my smartphone.

Has someone more ideas? If not I will have to contact the Plesk support ...

 

[Pleskie]

Maybe your host/server is experiencing down time?

 

[Nik G]

Goodfred said that the server has not been rebooted. So I don't think that it is downtime.
It can be overselled hardware node with heavy neighborhood VM's.

Having current info I have no ideas at this time.
Probably technical support will be better solution in this case - they have more abilities for deeper investigation (include logs inspecting, firewall configuration and so on).

 

[Pleskie]

It can be overselled hardware node with heavy neighborhood VM's.

I think we're refering to the same. I mean a cheap(er) host with none responsive VPS due to heavy traffic and too little processing power and RAM. But like you said it's just guessing. Your advice to contact technical support of the hosting company seems like the best advice to me.

 

[Goodfred]

I am trying a new fresh installation and I hope the problem wont occur on the new server.
If the problem is still there I will contact the Plesk Support and give you the solution if there is one.
Thanks!

 

[Pleskie]

I am trying a new fresh installation and I hope the problem wont occur on the new server.
If the problem is still there I will contact the Plesk Support and give you the solution if there is one.
Thanks!

If the problem is still there after a fresh install, you should probably contact technical support of your hosting company as @Nik G suggested earlier.

 

[Goodfred]

It is not a network issue.
With the new server (on the same ESXi) I dont have connection aborts.
It is the same as the old server (fresh Ubuntu 20.04, fresh Plesk)
At the moment Im doing a migration with the migration manager. (after fixing issues of the migration manager)(until now 0 connection errors/aborts)

 

[Nik G]

Actually Plesk does not change network's low-level settings or SSH daemon options.
So IMHO the problem is not Plesk-related.

btw, such situation also can be in case the same IP address assigned on more than two virtual servers.

Anyway, good to know that you was able to solve the problem, but pity that such radical way.