0

Ubuntu 14.04.1 LTS

apt-get update /upgrade was run last night at around 10pm. This morning people are saying that they can't get into ftp and wifi because these services are auth'd by freeradius.

Apt-get log: Start-Date: 2015-02-17 22:11:42 Commandline: apt-get upgrade Upgrade: python3-problem-report:amd64 (2.14.1-0ubuntu3.6, 2.14.1-0ubuntu3.7), php5-xmlrpc:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), php5-sqlite:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), libsystemd-login0:amd64 (204-5ubuntu20.9, 204-5ubuntu20.10), postfix:amd64 (2.11.0-1, 2.11.0-1ubuntu1), systemd-services:amd64 (204-5ubuntu20.9, 204-5ubuntu20.10), procps:amd64 (3.3.9-1ubuntu2, 3.3.9-1ubuntu2.2), php5-fpm:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), python-urllib3:amd64 (1.7.1-1build1, 1.7.1-1ubuntu0.1), php5-ldap:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), php5-common:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), python3-software-properties:amd64 (0.92.37.2, 0.92.37.3), libfreetype6:amd64 (2.5.2-1ubuntu2.2, 2.5.2-1ubuntu2.3), libsystemd-daemon0:amd64 (204-5ubuntu20.9, 204-5ubuntu20.10), libpam-systemd:amd64 (204-5ubuntu20.9, 204-5ubuntu20.10), php5-readline:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), libprocps3:amd64 (3.3.9-1ubuntu2, 3.3.9-1ubuntu2.2), udev:amd64 (204-5ubuntu20.9, 204-5ubuntu20.10), libudev1:amd64 (204-5ubuntu20.9, 204-5ubuntu20.10), php5-cli:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), unzip:amd64 (6.0-9ubuntu1.2, 6.0-9ubuntu1.3), php5-mysqlnd:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), python3-distupgrade:amd64 (0.220.6, 0.220.7), apport:amd64 (2.14.1-0ubuntu3.6, 2.14.1-0ubuntu3.7), python3-apport:amd64 (2.14.1-0ubuntu3.6, 2.14.1-0ubuntu3.7), software-properties-common:amd64 (0.92.37.2, 0.92.37.3), ubuntu-release-upgrader-core:amd64 (0.220.6, 0.220.7), linux-libc-dev:amd64 (3.13.0-44.73, 3.13.0-45.74), php5-gd:amd64 (5.5.9+dfsg-1ubuntu4.5, 5.5.9+dfsg-1ubuntu4.6), rsyslog:amd64 (7.4.4-1ubuntu2.3, 7.4.4-1ubuntu2.5) End-Date: 2015-02-17 22:12:20 /var/log/apt/history.log (END)

/var/log/freeradius Wed Feb 18 10:18:12 2015 : Info: Loaded virtual server Wed Feb 18 10:18:12 2015 : Info: Ready to process requests. Wed Feb 18 10:18:16 2015 : Error: rlm_perl: perl_embed:: module = /var/vhosts/it/radius_auth/radius_wifi_login.pl , func = authorize exit status= Can't call method "bind" on an undefined value at /var/vhosts/it/radius_auth/radius_wifi_login.pl line 212, line 751. Wed Feb 18 10:18:16 2015 : Auth: Invalid user: [meypau10] (from client uni-ventura port 0 cli EC3586AC80F6) Wed Feb 18 10:18:18 2015 : Error: rlm_perl: perl_embed:: module = /var/vhosts/it/radius_auth/radius_wifi_login.pl , func = authorize exit status= Can't call method "bind" on an undefined value at /var/vhosts/it/radius_auth/radius_wifi_login.pl line 212, line 751. Wed Feb 18 10:18:18 2015 : Auth: Invalid user: [meypau10] (from client uni-ventura port 0 cli EC3586AC80F6)

Line 212 from /var/vhosts/it/radius_auth/radius_wifi_login.pl: 

my $o_ldap = Net::LDAP->new($config{'database'}{'connection'}{'ldap'}{'dsn'});
$o_ldap->bind($config{'database'}{'connection'}{'ldap'}{'bind_dn'},
              'password' => $config{'database'}{'connection'}{'ldap'}{'password'});
if (!defined($o_ldap)) {
    # the user is allowed if we failed
    logToRadius('error', '[ldap] Could not bind to server.');
    return;
}

Is it possible that some of the php5 upgrades that took place last night caused freeradius or ldap to break? How could I troubleshoot this incident and resolve? I don't have a snapshot of the server, and rolling back would be impossible because the server is in use constantly. Would it be possible to roll-back these updates only? I've made backups of /etc/freeradius/* and /var/vhosts/it/radius_auth/

JasonC
  • 196
  • 6
  • 15
  • I think it's because the server cannot resolve to our domain controller. when I try to ping our domain controller by name, it doesn't resolve. – JasonC Feb 18 '15 at 18:59
  • Ok so we resolved this due to temp nameserver entries were not saved after a restart. In /etc/resolvconf/resolv.conf.d/base I added: nameserver 10.x.x.x nameserver 10.x.x.x then ran resolvconf -u to have it update Restarted services, and the name resolved and the services started successfully. – JasonC Feb 18 '15 at 19:15

1 Answers1

0

Ok so we resolved this due to temp nameserver entries were not saved after a restart. In /etc/resolvconf/resolv.conf.d/base I added: nameserver 10.x.x.x nameserver 10.x.x.x then ran resolvconf -u to have it update Restarted services, and the name resolved and the services started successfully.

JasonC
  • 196
  • 6
  • 15