14

I'm trying to go failed (either incorrect username, password, or both) on my server.

I changed /etc/ssh/sshd_config from

# Logging
SyslogFacility AUTH 
LogLevel INFO

to

# Logging
SyslogFacility AUTH 
LogLevel VERBOSE

and have since tried multiple ssh attempts with both existing and non-exisiting users with random passwords thus failing. When checking /var/log/auth.log nothing appears and it is entirely blank.

What am I missing? Does some other process need to also be install and running on my system? I'm running Ubuntu.

Any help or guidance on this matter is more than welcome.

Thanks

Improve this question
8
  • 1
    Did you restart sshd?
    – bonsaiviking
    Dec 3, 2012 at 22:26
  • 1
    What does your syslog configuration look like? This would probably be a file at /etc/syslog.conf or /etc/rsyslog.conf or /etc/rsyslog.d/*.conf
    – Stefan Lasiewski
    Dec 4, 2012 at 0:03
  • @StefanLasiewski the first 2 are empty and /etc/rsyslog.d/*.conf says "$AddUnixListenSocket /var/spool/postfix/dev/log"
    – edev.io
    Dec 4, 2012 at 0:12
  • @Georgejnr : If that is the case, it appears that the syslog configuration on your system is broken. There is normally a syslog file under /etc/syslog.conf or /etc/rsyslog.conf , and normally there should be more then one file under /etc/rsyslog.d/*.conf . Does ps aux show a syslog process?
    – Stefan Lasiewski
    Dec 4, 2012 at 0:33
  • @StefanLasiewski no it is not listed in ps aux. The previous sysadmin went a little rogue and broke a few things I believe on purpose. Think this could be part of it? How do I go about fixing this issue?
    – edev.io
    Dec 4, 2012 at 0:42

5 Answers 5

Reset to default
7

The LogLevel generally (apparently application dependent) refers to one of the defined severity levels supported by the system logging process (syslog). So change it back and restart the sshd server.

Now if you are not getting the output, you need to look at the system /etc/syslog.conf and see what MINIMUM loglevel the AUTH type of requests are being logged and to what file. The errors might be going to a different log file. OR you might not be logging these errors due to the syslog.conf configuration for the AUTH service. For more information consult the man pages on and syslog.conf.

Improve this answer
9
  • From sshd_config(5) LogLevel: Gives the verbosity level that is used when logging messages from sshd(8). The possible values are: QUIET, FATAL, ERROR, INFO, VERBOSE, DEBUG, DEBUG1, DEBUG2, and DEBUG3.
    – bonsaiviking
    Dec 3, 2012 at 22:28
  • 1
    my /syslog.conf is empty. I must add that I am taking over someone else's system and it seems that they didn't do a very good job of setting it up. Does the lack of syslog.conf mean that I am missing a service? (thanks for your response)
    – edev.io
    Dec 3, 2012 at 22:31
  • File is in /etc......it is possible that you might not be logging anything.
    – mdpc
    Dec 3, 2012 at 22:33
  • About VERBOSE in sshd_config....my mistake, but its not a syslog log level which is commonly asked for in many of the programs I have dealt with.
    – mdpc
    Dec 3, 2012 at 22:33
  • leaving VERBOSE still in my sshd_config and running sudo /etc/init.d/ssh restart it still isn't logging. Am I being dumb about something?
    – edev.io
    Dec 3, 2012 at 22:40
7

When I had the same problem on Debian, I found I had to restart rsyslogd:

/etc/init.d/rsyslog restart

(Your syslogd program may vary.)

The it started writing to /var/log/auth.log again.

Perhaps it had stopped logging after a disk full event, I'm not sure.

See also: https://bugs.launchpad.net/ubuntu/+source/rsyslog/+bug/1059854/comments/9

Improve this answer
1
  • 1
    This worked for me, but using systemctl instead to restart the syslog service (Debian sid using inetutils-syslogd). systemctl restart inetutils-syslogd.service
    – Brian Minton
    Sep 6, 2018 at 15:18
4

In my case the problem was with the ownership of the /var/log/auth.log file. It was owned by root:root but must be syslog:adm. Change with

sudo chown syslog:adm /var/log/auth.log

It appears to be a common problem with the newly created systems - there were more log files, which had this issue.

Improve this answer
3

In my case there was no diskspace on left on the root file-system /, which you can check for with df -h

Improve this answer
0
0

You could use a daemon like sshlog (I'm a contributor for this project) which monitors successful and failing ssh requests as well as the commands executed. The nice thing about this approach is that you can configure how you would like these events logged/alerted.

For example, you could configure it to report all failed login attempts and send that to a remote syslog, and send any sudo attempts to Slack alerts.

In this case the config would look like:

events:
  - event: log_all_failed_connections
    triggers:
      - connection_auth_failed
    actions:
      - action: syslog_to_localhost
        plugin: syslog_action
        server_address: remote_server_addr
        port: 514
        udp: True
        output_json: False
        program_name: sshlog

  - event: sudo_successful
    triggers: [ 'command_finish']
    filters:
      command_name: 'sudo'
      command_exit_code: '= 0'
    actions:
      - action: send_to_slack
        plugin: slack_action
        slack_webhook_url: 'https://hooks.slack.com/services/x/y/z'

More config examples are available here: https://github.com/sshlog/agent/tree/master/daemon/config_samples

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .