OpenSSH Troubleshooting: Difference between revisions
(4 intermediate revisions by the same user not shown) | |||
Line 2: | Line 2: | ||
* [[OpenSSH#Subjects|OpenSSH]] | * [[OpenSSH#Subjects|OpenSSH]] | ||
=Troubleshooting sshd Server= | =Troubleshooting sshd Server= | ||
Line 28: | Line 22: | ||
# update logging verbosity | # update logging verbosity | ||
systemctl start sshd | systemctl start sshd | ||
Increased log output will be available in /var/log/secure. | |||
=Scenarios= | |||
==Connection Takes a Long Time== | |||
In most cases this is caused by the server attempting to use DNS to get the IP name of the client. It is fixed by turning 'UseDNS' off. For more details, see: | |||
{{Internal|Sshd_Configuration#Turn_Off_Client_Name_DNS_Verification|Turn Off Client Name DNS Verification}} | |||
==Read from socket failed: Connection reset by peer== | |||
Symtoms: | |||
ssh user@ip fails with: | |||
NOMBP2:~ ovidiu$ ssh openshift@172.20.2.1 | |||
Read from socket failed: Connection reset by peer | |||
/var/log/secure shows: | |||
Sep 11 12:33:40 openshift-master1 sshd[1442]: fatal: No supported key exchange algorithms [preauth] | |||
Sep 11 12:35:36 openshift-master1 sshd[1445]: fatal: No supported key exchange algorithms [preauth] | |||
Root cause: | |||
The host key files had zero length. Not sure why is that, probably the installation procedure did not complete correctly. | |||
To fix: | |||
ssh-keygen -t rsa -f /etc/ssh/ssh_host_rsa_key | |||
ssh-keygen -t ecdsa -f /etc/ssh/ssh_host_ecdsa_key | |||
ssh-keygen -t ed25519 -f /etc/ssh/ssh_host_ed25519_key | |||
with empty passwords. | |||
==<tt>scp</tt>: subsystem request failed on channel 0== | |||
{{Internal|Scp#scp:_subsystem_request_failed_on_channel_0|scp | subsystem request failed on channel 0}} |
Latest revision as of 03:59, 25 December 2023
Internal
Troubleshooting sshd Server
Execute the sshd Server in Foreground
One method to troubleshoot a sshd server in an attempt to figure out why a specific ssh connection does not work as intended, is to stop the current sshd server (stopping the server won't drop the existing connections) and start it in foreground:
systemctl stop sshd /usr/sbin/sshd -d [-D] [-d] [-d]
sshd running in foreground behaves differently than the same binary running in background - for example, a foreground server will allow root logging with a public key without password, while the background server won't, so this method has limited usefulness in some cases.
Turn On Debug Logging
Stop the current sshd server (stopping the server won't drop the existing connections), increase logging verbosity as described here: sshd Logging Verbosity, and restart the server.
systemctl stop sshd # update logging verbosity systemctl start sshd
Increased log output will be available in /var/log/secure.
Scenarios
Connection Takes a Long Time
In most cases this is caused by the server attempting to use DNS to get the IP name of the client. It is fixed by turning 'UseDNS' off. For more details, see:
Read from socket failed: Connection reset by peer
Symtoms:
ssh user@ip fails with:
NOMBP2:~ ovidiu$ ssh openshift@172.20.2.1 Read from socket failed: Connection reset by peer
/var/log/secure shows:
Sep 11 12:33:40 openshift-master1 sshd[1442]: fatal: No supported key exchange algorithms [preauth] Sep 11 12:35:36 openshift-master1 sshd[1445]: fatal: No supported key exchange algorithms [preauth]
Root cause:
The host key files had zero length. Not sure why is that, probably the installation procedure did not complete correctly.
To fix:
ssh-keygen -t rsa -f /etc/ssh/ssh_host_rsa_key ssh-keygen -t ecdsa -f /etc/ssh/ssh_host_ecdsa_key ssh-keygen -t ed25519 -f /etc/ssh/ssh_host_ed25519_key
with empty passwords.