OVH OVH dedicated servers

Didn't even think about that. I'm reinstalling now. I'm going to test it out with the only change being the ListenAddress. See what happens.

These are the only changes I made to the stock system which led to the refuse connection:

Code:

nano /etc/ssh/sshd_config

RSAAuthentication yes
PubkeyAuthentication yes
PermitEmptyPasswords no
PasswordAuthentication no
ListenAddress <main.server.ip.address>

Obviously, added the authorized_keys file to /.ssh folder

 

Code:

grep 'listening' /var/log/secure
Apr 19 03:38:10 stock sshd[838]: Server listening on 0.0.0.0 port 22.
Apr 19 03:38:10 stock sshd[838]: Server listening on :: port 22.
Apr 19 03:40:53 stock sshd[11202]: Server listening on 0.0.0.0 port 22.
Apr 19 03:40:53 stock sshd[11202]: Server listening on :: port 22.
Apr 18 21:41:01 nsXXXXXX sshd[XXXXXX]: Server listening on 0.0.0.0 port 22.
Apr 18 21:41:01 nsXXXXXX sshd[XXXXXX]: Server listening on :: port 22.

I wonder if I have to leave the ListenAddress 0.0.0.0 instead of replacing the 0.0.0.0 with my server IP.

Code:

ListenAddress 0.0.0.0
ListenAddress <main.server.ip.address>

 

After adding

Code:

ListenAddress 0.0.0.0
ListenAddress <main.server.ip.address>

Doesn't look like the server IP is showing up.

Code:

grep 'listening' /var/log/secure
Apr 19 03:38:10 stock sshd[838]: Server listening on 0.0.0.0 port 22.
Apr 19 03:38:10 stock sshd[838]: Server listening on :: port 22.
Apr 19 03:40:53 stock sshd[11202]: Server listening on 0.0.0.0 port 22.
Apr 19 03:40:53 stock sshd[11202]: Server listening on :: port 22.
Apr 18 21:41:01 nsXXXXXX sshd[XXXXXX]: Server listening on 0.0.0.0 port 22.
Apr 18 21:41:01 nsXXXXXX sshd[XXXXXX]: Server listening on :: port 22.

UPDATE:
Removed ListenAddress 0.0.0.0 and rebooted. Tried to log back in and got connection refused. Logged in via IPMI and got the same results to the grep

Code:

grep 'listening' /var/log/secure
Apr 19 03:38:10 stock sshd[838]: Server listening on 0.0.0.0 port 22.
Apr 19 03:38:10 stock sshd[838]: Server listening on :: port 22.
Apr 19 03:40:53 stock sshd[11202]: Server listening on 0.0.0.0 port 22.
Apr 19 03:40:53 stock sshd[11202]: Server listening on :: port 22.
Apr 18 21:41:01 nsXXXXXX sshd[XXXXXX]: Server listening on 0.0.0.0 port 22.
Apr 18 21:41:01 nsXXXXXX sshd[XXXXXX]: Server listening on :: port 22.

UPDATE2
Figured it out. For some reason a reboot doesn't make the necessary changes to the sshd. Have to restart sshd after the changes are made.

 

I think I found the problem. For some reason when I reboot the machine, something is happening preventing me for logging in - "connection refused". Once I login through IPMI and restart sshd I am able to login. Not sure if this is a bug in the system or what.

Stock Centos 7 system.

 

Interesting result after boot.

Code:

# systemctl status sshd

Code:

sshd.service - OpenSSH server daemon
Loaded: loaded (/usr/lib/systemd/system/sshd.service;  enabled; vendor preset: enabled)
Active:  failed (Result: exit-code) since 11h ago
Docs:  man:sshd(8)
    man:sshd_config(5)
Process:  789 ExecStart=/usr/sbin/sshd -D $OPTIONS (code=exited, status=255)
Main PID:  789 (code=exited, status=255)

systemd[1]:  Stating OpenSSH server daemon...
sshd[789]: error:  Bind to port 22 on XXX.XXX.XXX.XXX failed:  Cannot assign requested address.
sshd[789]: fatal:  Cannot bind any address.
systemd[1]: sshd.service:  main process exited, code=exited, status=255/n/a
systemd[1]: Failed to start OpenSSH server daemon.
systemd[1]: Unit sshd.service entered failed state.
systemd[1]: sshd.service failed.

Login via IPMI and restart sshd via service sshd restart.

Code:

# systemctl status sshd

Everything is fine and the IP is listening on XXX.XXX.XXX.XXX port 22.

So, after reboot I get the error and have to restart sshd to get it working again.

 

I'm using the main server IP address. All that I've done to this box is add nano and edit the sshd_config file. The only interface on this machine is eth0 and is populated by OVH on install.

I just checked eth0 and it's setup correctly for being the main IP. This should work fine... or I'm going crazy... one of the two.

 

Dam ListenAddress! Thanks for searching for / finding that bug information. I was searching last night for an issue and couldn't find anything about using ListenAddress.

Yea, going to skip this for now so I can move onto actually getting the server setup.

 

Can you go into detail on what you needed to do? Looking at a new one with NVMe and want to use original kernel.

 

I was having an issue in the beginning, but it appears that they must have modified their centos image. I've reinstalled several times and haven't had an issue since using the org kernel. You should be good to go.

Note: I'm using the 1.2TB NVMe, so I know everything works with that model.

Worst comes to worst, you can always install stock centos from an iso in the IPMI, you just have to attach a virtual device linked to the iso on your desktop.

 

What at the reasons to use the original kernel rather than ovh kernel? Ovh kernel should be more tested in OVH servers. I think I've read @eva2000 were recommended Linode kernel for linode servers

 

@eva2000 recommends using the org kernel on OVH. I believe there is a warning notice if you try to install CMM using the OVH kernel. @eva2000 would know more why.