DevHeads.net

Postings by Andreas Hasenack

samba 4.9 in disco

Hi,

I just wanted to email the list explaining why samba 4.9 is not yet in
disco. It's also blocking the ldb migration.

The samba DEP8 tests pass, but one of the triggered tests (freeipa)
actually showed what we think is a valid bug, or at least significant
change in behavior, in samba[1].

Basically, in a fresh install, if you have winbind running, smbd won't
start, even in standalone mode (not part of a domain). This happens in
debian as well, fedora, and possibly suse. There are several mailing
list threads[2,3,4] about it.

The freeipa dep8 tests caught it by accident really.

help with sbuild dep8 failures on i386/bionic

Hi,

I have a couple of SRUs that are triggering sbuild DEP8 runs on bionic
and these are only failing on i386:

<a href="http://autopkgtest.ubuntu.com/packages/s/sbuild/bionic/i386" title="http://autopkgtest.ubuntu.com/packages/s/sbuild/bionic/i386">http://autopkgtest.ubuntu.com/packages/s/sbuild/bionic/i386</a>
<a href="http://autopkgtest.ubuntu.com/packages/s/sbuild/bionic/amd64" title="http://autopkgtest.ubuntu.com/packages/s/sbuild/bionic/amd64">http://autopkgtest.ubuntu.com/packages/s/sbuild/bionic/amd64</a>

I triggered those using debootstrap from proposed, otherwise we hit
another bug where "disco" is unknown.

I can reproduce the failure on bionic i386, but I cannot explain how
it is even supposed to work in any architecture.

The test basically uses sbuild to build procenv on the current devel
version of ubuntu. Today, that's disco.

server seed: remove acpid?

Hi,

we are reviewing the server seeds
(<a href="https://code.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/+git/ubuntu/+ref/disco" title="https://code.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/+git/ubuntu/+ref/disco">https://code.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/+git/ubuntu/+re...</a>)
and the acpid package came up.

python3-numpy depending on *both* python 3.6 and 3.7

Hi,

while investigating some DEP8 failures currently in cosmic's
migration, I came across this:
$ dpkg -s python3-numpy|grep Depends
Depends: python3 (<< 3.8), python3 (>= 3.6~), python3.6:any,
python3.7:any, python3:any (>= 3.3.2-2~), libblas3 | libblas.so.3,
libc6 (>= 2.27), liblapack3 | liblapack.so.3

Is it ok/correct to depend on two python versions like that? Is the
point of it making sure numpy is available regardless which python 3
you are using?

Proper way to handle britney hints

Hi,

I would like to know how to properly handle updates to the so called
"britney hints", which controls how DEP8 test failures are handled in
migrations. Particularly for the development release.

I uploaded <a href="https://launchpad.net/ubuntu/+source/ocfs2-tools/1.8.5-5ubuntu1" title="https://launchpad.net/ubuntu/+source/ocfs2-tools/1.8.5-5ubuntu1">https://launchpad.net/ubuntu/+source/ocfs2-tools/1.8.5-5ubuntu1</a>
on Jul 31st, and it has a known failure[1][2] on s390x, for which I
proposed this branch[2]. I also pinged in #ubuntu-release several
times after that.

It seems this communication got lost somehow.

-Wl,-Bsymbolic-functions in Ubuntu but not debian/ why?

Hi,

does anybody recall why Ubuntu has -Wl,-Bsymbolic-functions in
dpkg-buildflags --get LDFLAGS but not Debian?

NULL pointer deref in pcf_check_dbms_client() with unreadable map file

Hi,

postfix-3.3.0

we got a bug report (<a href="https://bugs.launchpad.net/ubuntu/+source/postfix/+bug/" title="https://bugs.launchpad.net/ubuntu/+source/postfix/+bug/">https://bugs.launchpad.net/ubuntu/+source/postfix/+bug/</a>
1753470) where postconf was crashing if main.cf had a map pointing to a
file that the user couldn't read.

ubuntu@bionic-postfix:~$ l /etc/postfix/valiases.cf
-rw-r----- 1 root root 169 May 7 14:08 /etc/postfix/valiases.cf

ubuntu@bionic-postfix:~$ cat /etc/postfix/valiases.cf
cat: /etc/postfix/valiases.cf: Permission denied

ubuntu@bionic-postfix:~$ postconf
Segmentation fault (core dumped)

ubuntu@bionic-postfix:~$

gdb shows the crash is in vstream_fileno(fp):
Program received signal SIGSEGV, S

hidden symbol `pthread_atfork' in /usr/lib/x86_64-linux-gnu/libpthread_nonshared.a(pthread_atfork.oS) is referenced by DSO

Hi,

yesterday evening, and today, I upgraded a bionic container where I was
building the nvml packages, and to my surprise a build that worked last
thursday stopped working today:

make[3]: Entering directory
'/home/ubuntu/nvdimm/nvml/nvml-git/src/examples/librpmem'
cc -o manpage manpage.o -Wl,-rpath=../../debug -L../../debug -lrpmem
-pthread -lfabric
/usr/bin/x86_64-linux-gnu-ld: manpage: hidden symbol `pthread_atfork' in
/usr/lib/x86_64-linux-gnu/libpthread_nonshared.a(pthread_atfork.oS) is
referenced by DSO
/usr/bin/x86_64-linux-gnu-ld: final link failed: Bad value
collect2: error: ld ret

"testbed out of date"

Hi,

I'm getting this error a lot in some test runs (like
<a href="https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-artful/artful/s390x/f/firejail/20171207_130707_4ca6a@/log.gz" title="https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-artful/artful/s390x/f/firejail/20171207_130707_4ca6a@/log.gz">https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a8492...</a>
):

autopkgtest [13:07:06]: ERROR: erroneous package: Test dependencies
are unsatisfiable. A common reason is that your testbed is out of date
with respect to the archive, and you need to use a current testbed or
run apt-get update or use -U.
blame: firejail
badpkg: Test dependencies are unsatisfiable.

"sponsorhip for retrying dep8 tests"

Hi,

many of you probably have been in a similar situation: you take your time
to fix a bug in package foo, upload it, and then the migration is blocked
because the dep8 tests of package bar failed. You go and check, and it's an
error in something completely unrelated to what you changed in your upload,
or a test infrastructure problem. You investigate the history of that test,
see a few cases of the same exact error, and trigger a retry. Then it
works, and you move on.

Except that when your upload was sponsored, you can't click retry.

Server team meeting minutes: 2017-09-12

== Meeting information ==
* #ubuntu-meeting: ubuntu-server-team, 12 Sep at 16:00 &mdash; 16:19 UTC
* Full logs at
[[http://ubottu.com/meetingology/logs/ubuntu-meeting/2017/ubuntu-meeting.2017-09-12-16.00.log.html]]

== Meeting summary ==

=== Review ACTION points from previous meeting ===
The discussion about "Review ACTION points from previous meeting"
started at 16:00.

* ''ACTION:'' nacc to write a release notes entry on ipv6 netboot
(carried over)
* ''ACTION:'' nacc to write a server guide entry on ipv6 netboot
(carried over)
* ''ACTION:'' rbasak to add maintainership info to mys

Server team meeting minutes: 2017-06-27

== Meeting information ==
* #ubuntu-meeting: ubuntu-server-team, 27 Jun at 16:00 &mdash; 16:40 UTC
* Full logs at [[
<a href="http://ubottu.com/meetingology/logs/ubuntu-meeting/2017/ubuntu-meeting.2017-06-27-16.00.log.html" title="http://ubottu.com/meetingology/logs/ubuntu-meeting/2017/ubuntu-meeting.2017-06-27-16.00.log.html">http://ubottu.com/meetingology/logs/ubuntu-meeting/2017/ubuntu-meeting.2...</a>
]]

== Meeting summary ==

=== Review ACTION points from previous meeting ===
The discussion about "Review ACTION points from previous meeting" started
at 16:00.

* ''LINK:''
<a href="https://lists.ubuntu.com/archives/ubuntu-devel/2017-June/039831.html" title="https://lists.ubuntu.com/archives/ubuntu-devel/2017-June/039831.html">https://lists.ubuntu.com/archives/ubuntu-devel/2017-June/039831.html</a>
* ''ACTION:'' * nacc to write a release notes entry on ipv6 netboot
(carried over)
* ''ACTION:'' * nacc to write a server guide entr

RFC: disabling the SMB1 protocol in our samba stack (client, server) (#1697817)

Hi,

we had a bug [1] filed against Ubuntu as part of a larger campaign [2] to
disable the old, vulnerable, deprecated SMB1 protocol in the Samba stack.

I asked upstream [3] for their opinion on changing the default protocol
level.

In terms of an actual change, it's just a pair of independent configuration
settings in smb.conf (if upstream doesn't change the defaults): one for
client, one for server. That part is simple. The repercussions need to be
thought through, however.

RFC: disabling the SMB1 protocol in our samba stack (client, server) (#1697817)

Hi,

we had a bug [1] filed against Ubuntu as part of a larger campaign [2] to
disable the old, vulnerable, deprecated SMB1 protocol in the Samba stack.

I asked upstream [3] for their opinion on changing the default protocol
level.

In terms of an actual change, it's just a pair of independent configuration
settings in smb.conf (if upstream doesn't change the defaults): one for
client, one for server. That part is simple. The repercussions need to be
thought through, however.

Should gir1.2-gudev-1.0 depend on libgudev-1.0?

Apologies if this was already discussed in this mailing list.

I got a bug filed against landscape-client today
(<a href="https://bugs.launchpad.net/landscape-client/+bug/1159997" title="https://bugs.launchpad.net/landscape-client/+bug/1159997">https://bugs.launchpad.net/landscape-client/+bug/1159997</a>) complaining
that it wouldn't start on Raring unless libgudev-1.0-0 is installed.

I checked and it seems that on Oneiric, Precise and Quantal
gir1.2-gudev-1.0 pulls in libgudev-1.0, but not on Raring.

Could PPAs accept <distro>-proposed pockets?

Hi,

I tend to use PPAs to test-build my packages before I upload to
proposed for SRUs.

PPAs are very convenient for this. I get to test the build and it also
makes the packages available for me in a repository.

Problem is, I tend to forget to adjust the changelog line from, say,
"precise-proposed" to "precise" before uploading to the PPA, and then
I get this rejection email:

Rejected:
PPA uploads must be for the RELEASE pocket.

Would it be possible to make the PPAs accept these <distro>-proposed
uploads? Maybe mangle them into <distro> on-the-fly or something?

Packaging a twistd plugin and python-support issues

Hi,

I'm packaging a twisted plugin (<a href="https://launchpad.net/txstatsd" title="https://launchpad.net/txstatsd">https://launchpad.net/txstatsd</a>) for
Lucid (so no dh_python2 for me) and ran into an issue with
python-support that I would like some advice about.

If I don't create a ".noinit" file in the twisted plugin directory,
python-support will create an empty __init__.py file in there in
postinst.