Postings by Robert Heller

CentOS 6 SELinux question: inbound ssh.

OK, after beating my head against the wall for an hour or so, I finally
figured out why I could not ssh from a MacMini (running MacOSX 10.11.6) to my
Linux Desktop (running CentOS 6), using the amandabackup account with public
key authentification.

Ubuntu 18.04 GUI File Manager Weirdness

OK, first of all *I* don't use a GUI File Manager. However I provide IT
support for the local library where most people do use the GUI File Manager.
All of the machines are running an up-to-date Ubuntu 18.04. The only other
weirdness is that the /home filesystem is NFS mounted and user authentifcation
is via LDAP.

What is happening is that users cannot open folders on their desktop. Either
left clicking on the folder icon or right-click and Open. Nothing happens.

Alternitives to Firefox...

OK, I recently ugraded to the current ESR release of Firefox for CentOS 6.
And I am having problems with the user interface (basically it has become hard
[for me] to use).

What alternitives are there? (Chrome and Chromium are not possible with
CentOS, and Chrome and Chromium are actually worse).

Debmirror: *remove* a distro version

I current have a local mirror that is maintained with a daily run of
debmirror. Right now, three versions of Ubuntu are maintained: 14.04, 16.04,
and 18.04. Sometime soon, I am going to drop 14.04. Because of the way the the
Ubuntu mirror is structured (all of the package files for all three versions
are in a single tree of package files), I am not sure how to remove just the
(older) 14.04 versions. With CentOS, this is easy, since its repo is
structured having a separate tree for each major version and removing version
is just a single rm -rf command for that verison's tree.

Ubuntu 18.04: systemd-resolved -- crashing or failing to start properly...

I have an *intermittent* problem with systemd-resolved. It is either crashing
or fails to start properly. The sympton is that when I go to do an update (eg
apt-get update), I get name resolver errors to the *local* (LAN only) mirror
host. Doing a "systemctl restart systemd-resolved" cures the problem. I am not
sure exactly what is going on, but it appears that for some reason
systemd-resolved is not seeing (or using) the *local* LAN DNS server (bind9
running on a CentOS 6 server).

Wireless not connecting...

I just installed 18.04 on an older Hp Laptop. Everything is working, except
(of course!) the WiFi.

The WiFi NIC is a Realtek Semiconductor Co.

CentOS 6/EPel: Missing EPel package?

The EPel repo has a *nearly* complete collection of QT5 packages. One
important one that *seems* to be missing: qmake.

There does not seem to be a package containing qmake in the collection of QT5
packages for CentOS 6!

I guess I can find the source code and build it (somehow)...

GDM3 Login List -- where does this list come from?

I manage a batch of diskless workstations and have *automatic* login
(AutomaticLogin options in /etc/gdm3/custom.conf) set up (for a guest
account). *Sometimes* people need to login with a specific personal account.
What I would like would be for the *automatic* login to have a delay. On
machines *with* disks, this can be done with the TimedLogin options in
/etc/gdm3/custom.conf. The problem is this: the logins are authenticated via

"Untrusted application launcher (desktop launchers)"

I am having this problem on Ubuntu 18.04 -- I manage a batch of desktop
machines with some convience desktop launchers, which gnome3 insists are
"untrusted". With some general websearching reveals that this is a *GNome3*
so-called "security" issue
(<a href="" title=""></a>).

18.04: "untrusted launcher" -- how to make it *permanently* trusted

I recently upgraded a batch of diskless workstations to Ubuntu 14.04 using
DRBL. Everything is working well, except for one small problem: We have some
launcher files (.desktop) files in the login account Desktop folder, but even
though they are chmod +x (and some are in fact direct copies from
/usr/share/applications) they come up as "untrusted". (And yes, they are
properly owned by the user and group, etc.). They work, once you click "Trust
and execute", but they become untrusted again on reboot.

What is going on here? Is there some additional magic needed somewhere?

do-release-upgrade is crashing when going from 16.04 to 18.04

I recently upgraded a laptop from 14.04 to 16.04 using do-release-upgrade. I
has been running just fine with 16.04, but I just tried to update it to 18.04
by running do-release-upgrade again, but do-release-upgrade is crashing.

The only *possible* weirdness is that I am using a local mirror (built with
debmirror). I did use this mirror to install a 18.04 VM (and have been
successfully keeping the VM up-to-date). And I did use the mirror to upgrade
this laptop, another laptop, and (so far) one desktop machine from 14.04 to

What should I be looking for to debug this problem?

Getting slapd under CentOS 6 to log connections, etc.

I am running the stock OpenLDAP-servers on a CentOS 6 system. Everything is
working just fine, but I am trying to debug connections from a Ubuntu (18.04)
system and want to turn on logging in slapd, but although I can get slapd to
create a log file, it never actually writes anything.

I have these two lines in /etc/openldap/slapd.d/cn=config.ldif:

olcLogFile: /var/log/slapd/log
olcLogLevel: conns BER config ACL stats none

And although /var/log/slapd/log gets created, nothing is ever written to it.

What am I missing?

"Authentication failure" on Ubuntu 18.04...

I have a server running CentOS 6 on the bare metal, running the stock CentOS 6
OpenLDAP packages (openldap-2.4.40-16.el6.x86_64,
openldap-clients-2.4.40-16.el6.x86_64, openldap-2.4.40-16.el6.i686,
openldap-servers-2.4.40-16.el6.x86_64) [yes I know these are old, but they
*work* -- "If it ain't broke, don't fix it."]. There are two VMs running on
this server, one running Ubuntu 14.04 and the other Ubuntu 18.04.

Ubuntu 18.04 as LDAP Authentification Client -- What am I missing?

OK, I followed the instructions here:

<a href="" title=""></a>

But things are not working:

Script started on Wed 06 Mar 2019 04:38:06 PM EST slogin ub180464

Welcome to Ubuntu 18.04.2 LTS (GNU/Linux 4.15.0-46-generic x86_64)

* Documentation: <a href="" title=""></a>
* Management: <a href="" title=""></a>
* Support: <a href="" title=""></a>

* Canonical Livepatch is available for installation.
- Reduce system reboots and improve kernel security.

MATE questions

Is there a mailing list for the MATE desktop? I am not finding it with

I am some configuration questions relating to MATE.

Questions about Ubuntu 18.04 live/install ISO

I am trying to run/install Ubuntu 18.04 in a VM on my CentOS 6 Laptop and it
is never coming up. The boot process seems to be looping. I have two

Does live/install ISO Ubuntu 18.04 need a connection to the public Internet to
come up? I only have dialup Internet and initially, the network will be to a
closed vlan. Is the a problem?

Also, I have only allocated 1G of RAM to the VM. Is this too little?

CentOS 6: Logrotate / selinux problem

Ever since a recent power failure I have been getting a Logrotate error. My
machine is on a UPS -- it shutdown cleanly, but I suspect that its BIOS/RTC
battery is dead, since the machine came up thinking it was 1982 :-(. I reset
the clock and everything is fine, *except* I had to delete Logrotate's state
files (which had bad dates).

Ubuntu 14.04 apt-get problem...

Amoungst a number of Ubuntu machines I manage at the local library are a pair
of laptops. Back a few kernel updates ago, there was a "bad" kernel (caused
Java to crash). Since these laptops have the Arduino IDE installed, this was
a problem. I could not figure out how to make them boot the previous kernel
by default (the grub2 config process basically sucks and Ubuntu 14.04
incarnation of it is basically broken in this regard, but that is neither here
nor there), so I just did an apt-get purge of the "bad" kernel.

Failed to get to installer for CentOS 7 VM under CentOS 6...

I just tried to create a CentOS 7 VM on a CentOS 6 host, but it crashes to the
dracut prompt. I am using the PXEBoot installer. I *think* it is unhappy
with the (virtual) graphics controller, but I am not sure.

The rdsosreport.txt file is available here:
<a href="" title=""></a>

"The system detected a problem, do you want to report it?" dialog

I'm getting this "The system detected a problem, do you want to report it?"
dialog. It does not say what the problem is. How do I find out what the
problem is? I am not interested in reporting it. I suspect it is something
stupid that I can fix, but I cannot figure out how to find out what is wrong.
This is with Ubuntu 14.04 LTS. I am used to CentOS and used to a less
pointy-clicky user interface, where one gets real error messages, not these
silly "idiot light" dianostics.

An observation (kernel 3.13.0-155 is broken WRT Java) and a Grub2 / Ubuntu question

OK, I just did updates at the local library, where the machines (except the
server's bare metal) run Ubuntu 14.04 LTS. The kernel update (3.13.0-155)
crashes (sort of) when you try to run a Java program (specificly the Arduino
IDE (1.8.5)).

My question is this: what is the "magic" to change the default kernel to boot?
With the grub2 configuration "magic" it is not a simple matter of changing one
line /boot/grub/grub.conf, but seems to require something far more arcane and
there does not seem to be a simple man page that explains this.

Imap daemons for CentOS 6 (other then cyrus-imapd)

Are there any imap daemons (besides cyrus-imapd). cyrus-imapd is appearently
not compatible with postfix + procmail. I need an imap daemon that will work
with a postfix + procmail system.

Ubuntu stock 14.04 Firefox crashes on machines with NFS mounted file systems

We have 9 workstations, running Ubuntu 14.04, that are booted diskless using
drbl. For the past few versions of Firefox, (since 58 0r 59) Firefox crashes
on startup. It is basically not usable, even with a fresh profile. FF
52.6.0ESR does work however. Oh, Chrome and Chromium don't work either (never
have). We have a pair of laptops that also have Ubuntu 14.04 on them, but
they are setup to use their own internal disks. FF works on them. It
*appears* to be some weirdness with NFS (I think), but I don't know what.

Mailing list config question / minor rant

The <a href="mailto: ... at centos dot org"> ... at centos dot org</a> seems to be *uniquely* configured to hold messages with
a (single) CC to self. ALL of the other mailing lists I am on have no problem
with a CC to self. My mailer software is configured to CC to myself. This is
*only* a problem with the CentOS list. Why is this? Yes, I know it is *normal*
to configure Mailman to hold messages with a *long* CC list, since that is a
common tactic by spammers. But a CC of one? Really? Do spammers bother with
CC's with only one address?

CentOS6: HELP! EFI boot fails after replacing disks...

OK, I wanted to replace the 500G disks in a Dell T20 server with new 2TB
disks. The machine has 4 SATA ports, one used for the optical disk and three
for the hard drives.

Ubuntu 14.04's Debmirror broken?

I have been using debmirror (1:2.16ubuntu1.1) on an Ubuntu 14.04 server to
maintain a local mirror of Ubuntu 14.04, but in the past week it is failing:

Use of uninitialized value $l in chop at /usr/share/perl5/LockFile/ line 438.
Use of uninitialized value $l in string eq at /usr/share/perl5/LockFile/ line 439.
Use of uninitialized value $stamp in chop at /usr/share/perl5/LockFile/ line 570.
Use of uninitialized value $stamp in pattern match (m//) at /usr/share/perl5/LockFile/ line 576.
Use of uninitialized value $hostname in string ne at /usr/share/perl5/

How do I get the kernel srpm?

I tried to follow the work flow shown in <a href="" title=""></a>, but
it does not seem to work:

I did this:

mkdir CentOS
pushd CentOS
git clone <a href="" title=""></a>
git clone <a href="" title=""></a>
pushd kernel/
git checkout c6

And I got the message:

Missing metadata. Please run from inside a sources git repo

What am I missing?

Question about "generic" USB <=> Serial Port kernel drivers...

OK, I am running CentOS 6 on my Laptop and Desktop, presently kernel version
2.6.32-696.20.1.el6.x86_64 (yes, a version or so out of date -- I'll probably
do an update soon -- non-trivial since my Internet connection is via dialup).

I recently bought a couple of Adafruit Arduino compatibles, a Flora and a
Metro 328. Hooking up the Flora and the Metro to a Raspberry Pi running
Raspbian (kernel 4.9.35-v7+), tells me that the Metro 328's USB uses a FTDI
USB serial interface (ftdi_sio.ko) and the Flora uses the cdc_acm (cdc-acm.ko)
USB serial interface.

Posting check

I am wondering if I am able to post to the CentOS list.

Resetting and *locking* the Desktop in Ubuntu 14.04

We have a batch of public workstations and sometimes, somehow (I don't know
how -- I am not a user of pointy-clicky desktop file managers), the desktop
file manager displays ~home/ instead of ~home/Desktop. How do I fix this and
how do I prevent users from randomly changing it? Specificly, what config
files are being messed with and how do I "lock" those files (eg make them
non-writable or something). And no, I don't want a pointy-clicky solution, I
want a solution I can implement from the command line, since I suspect it
might be a matter of running a daily cron job to fix/re-fix.