Bug#997811: unbound crashes, remote dos?

[Stephane Tranchemer]

So, I took the source from Unbound main site for version 1.15 and made a 
package out of it using the deb-source.


Guess what, it is now running for a week and we didn't see a single crash.

Just my 2 cents but I think that Debian maintainers should consider 
going for better than 1.13 now.

Bug#948553: Tomcat9 startup script ignores CATALINA_PID

[Stephane Tranchemer]

Package: tomcat9
Version: 9.0.16-4

In debian 10 (buster)

/usr/share/tomcat9/bin/catalina.sh is ignoring the CATALINA_PID
variable therefore does not generate a PID file accordingly.

example startup command :

set -a; JAVA_HOME="/usr/lib/jvm/java-11-openjdk-amd64"; source
"/etc/default/tomcat9"; CATALINA_HOME="/usr/share/tomcat9";
CATALINA_BASE="/var/lib/tomcat9"; JAVA_OPTS="-Djava.awt.headless=true
-server -Xmx1280M -Xms1280M -Xss2M -Dfile.encoding=EUC-JP
-XX:+PrintGCDetails -XX:+HeapDumpOnOutOfMemoryError
-XX:HeapDumpPath=/volatile/dump/db2-11-test2 -Dcom.atson.test=true
-Dcom.atson.local=false -Dcom.atson.webapp=true
-Dcom.atson.useDB2Type4=false -Dcom.atson.useDataSourceResource=true
-Dcom.atson.XLINK_VIA_GW=true"; \ CATALINA_PID="/var/run/tomcat9.pid";
CATALINA_TMPDIR="/tmp/tomcat9-tomcat9-tmp"; LANG="C.UTF-8";
JSSE_HOME=""; cd "/var/lib/tomcat9";
"/usr/share/tomcat9/bin/catalina.sh" start

The origin of the problem can be found in this line (actually 2 identical
lines) :

2\>\&1 \&\& echo \$! \>\"$catalina_pid_file\" \; \} $catalina_out_command "&"

with this change it works as expected :

-2\>\&1 \&\& echo \$! \>\"$catalina_pid_file\" \; \}
$catalina_out_command "&" +2\>\&1 \& echo \$! \>\"$catalina_pid_file\"
\; \} $catalina_out_command "&"

Bug#909643: closed by Ben Hutchings (Re: doing a dist-upgrade from debian7 to debian8 ended with a non bootable system)

[Stephane Tranchemer]

Hi,
You are right Ben, this was the culprit.

The bottom line is, in a previous version of the debian distribution the
busybox package was called "busybox-cvs-static".

But at a time it just gone to "busybox-static", so when doing the normal
procedure of
apt-update
apt-upgrade
apt-distupgrade

it is not replaced by the newer version...

Quite a nasty pitfall.

Thanks for your time.


> The screenshot shows:
>
> "busybox 20040623-1"
>
> You have an ancient version of busybox installed and you must upgrade
> it.
>
> Ben.
>
> --
> Ben Hutchings
> Horngren's Observation:
>   Among economists, the real world is often a special case.
>
>
>

Bug#909643: doing a dist-upgrade from debian7 to debian8 ended with a non bootable system

[Stephane Tranchemer]

I did test once again with a different machine, same result.

It seems really that initramfs is the culprit here, the old 3.2 kernel from
debian7 boots fine.

when thrown back in busybox with the (initramfs) prompt you can do vgchange
-ay and access partitions.
So there's no valid reason to get an error about /etc/fstab being not
accessible nor readable.

Bug#909643: installation-reports: doing a dist-upgrade from debian7 to debian8 ended with a non bootable system

[Stephane Tranchemer]

Hello,

Thanks for your concern.


On 27 September 2018 at 10:32, Ben Hutchings  wrote:

>
> What is the kernel command line?  (Run "cat /proc/cmdline" to show it.)
>

With the bootable kernel 3.2

# uname -a
Linux ahs-clone 3.2.0-6-686-pae #1 SMP Debian 3.2.102-1 i686 GNU/Linux
# cat /etc/debian_version
8.11
# cat /proc/cmdline
BOOT_IMAGE=/vmlinuz-3.2.0-6-686-pae root=/dev/mapper/vg00-root ro

In a non-bootable state in initramfs prompt with kernel 3.16

BOOT_IMAGE=/vmlinuz-3.16.0-6-686-pae root=/dev/mapper/vg00-root ro


>
> Please also send a copy of the /etc/initramfs-tools/initramfs.conf
> file.
>

Here is the file in attachment.

Thank you.
-- 
-
Tranchemer Stephane
System dept.
Asahinet, inc.
Kabukiza tower 21st Fl.
4-12-15 Ginza
Chuo-ku, Tokyo
104-0061 Japan


initramfs.conf
Description: Binary data