Postfix email suddenly sloowww...

Discussion in 'General' started by SamTzu, Jan 14, 2016.

  1. SamTzu

    SamTzu Active Member

    Since yesterday ISPconfig3 postfix send mail suddenly became really slow. Squirrel mail send button seemed to get "stuck" for a minute.
    Every email takes like a minute to send but there does not seem to be anything wrong with the server.
    After reboot it's back to normal again.

    Any ideas?
     
  2. SamTzu

    SamTzu Active Member

    I'm thinking this is somehow related to recent update. Maybe on the hardware host level.
     
  3. till

    till Super Moderator Staff Member ISPConfig Developer

    Just to be sure, check with:

    postqueue -p

    that your server has not a full mail queue.
     
  4. SamTzu

    SamTzu Active Member

    Nope. Normal. Seems like this might be something to do with Proxmox OpenVZ. Some containers stop Really slow.
     
  5. SamTzu

    SamTzu Active Member

    Well. That did not help for long. The problem is back.
     
  6. till

    till Super Moderator Staff Member ISPConfig Developer

    Do you have high load (top command) in the affected vm's or on the host? If yes, which processes consume all the cpu cycles?
     
  7. SamTzu

    SamTzu Active Member

    Everything looks normal. Except these...
    Code:
    2016-01-14 16:20:33,970 fail2ban.actions: WARNING [ssh] Ban 59.45.79.55
    2016-01-14 16:24:06,056 fail2ban.actions: WARNING [sasl] Ban 24.213.96.31
    2016-01-14 16:31:32,632 fail2ban.actions: WARNING [ssh-ddos] Ban 94.102.53.197
    
    I'm thinking there is something vulnerable with SSH. Seems like there are patches for SSH upgrade.
     
  8. SamTzu

    SamTzu Active Member

    Told U.
    Malicious OpenSSH servers can silently steal people's private SSH keys as they try to login, it emerged today.
     
  9. SamTzu

    SamTzu Active Member

    And it's back again... even though I blocked SSH port. Looks like the ssh-ddos was not the only problem with this email server. Something else is also going on. The plot thickens :)
     
  10. SamTzu

    SamTzu Active Member

    I think I figured it out.
    It is the rsyslog update from last week. Default config works fine but all servers that tried to send their logs to central logging server slowly froze. Also there was problems with mysql-server update on another machine. Unattended upgrade failed to 5.6 without a reboot in between. Quality updates from Debian.
     
    till likes this.

Share This Page