SOLVED occasionally getting Error reading SSH protocol banner[Errno 54] Connection reset by peer..

[heupink]

Hi,
I setup replication between two TrueNAS instances, on the same subnet, no firewall in between, and we are getting occasional errors like:

Code:

Error reading SSH protocol banner[Errno 54] Connection reset by peer..

and

Code:

 failed: kex_exchange_identification: Connection closed by remote host
Broken pipe..

These errors are logged on the SENDING server. Nothing much particular is logged on the receiving side. (in /var/log/*)

It happens occasionally, and most of the time, replication works fine. This is all on latest truenas: Truenas-12.0-STABLE on both servers involved.

I only have defined one ssh connection on the sending server, and it is used for all replication. Usually starting the replication task manually (without making any changes) works immediately.

I would like to increase (dramatically) the log level for ssh on the receiving side, as itmight reveil something. However...increasing of the logging seems to be for SFTP, and nothing more is logged for SSH.

How to troubleshoot this?

 

[heupink]

I have now added

Code:

MaxStartups=100

to ssh config. Let's see if that helps.

 

[heupink]

It's thursday now, and we have not seen the error anymore after adding the above sshd config on monday. I guess everybody with multiple replication jobs through ssh should increase the MaxStartups.

 

[saurav]

I upgraded from 11.3 to 12.0-U8, and this seems to be working for me as well.