[opensuse-factory] openSUSE Tumbleweed - Review of the weeks 2016/6 & 7
Dear Tumbleweed users and hackers, Again I cover two weeks, with the 2nd one being more of troubles than releases. The snapshots covered this week are 0205, 0207, 0208, 0209 (week 6) and 0212 (week 7). I’ll try to explain at the end what was holding us back in week 7. Noteworthy updates as part of the last snapshots: * AppStream-glib update. Most notably: the icons are no longer duplicated across multiple repositories metadata * GCC 5.3.1 is now the default compiler * Mozilla Firefox 44 * cmake 3.4.3 – most importantly, ‘modules’ no longer get a soname during linking. Some packages might fail rpmlint now in case the shared library packaging policy was not applied * GStreamer 1.6.3 – Many bugfixes * Plasma 5.5.4 – you know you want it! What is brewing in the next snapshot (0217+) * Firewalld – available for testers. A future version should bring migration assistance from SUSEFirewall2 to firewalld * plymouth gained a new plugin to show simple texts (used to indicate which partition to unlock) I would like to introduce a new section, mentioning ‘bugs I feel are noteworthy’. The list is incomplete and might not cover the bug YOU had seen, and it shall not be a complete list of new bugs reported. * Tumbleweed users using LUKS encrypted devices reported plymouth related issues. Current workaround identified is to disable plymouth at boot time with “plymouth.enable=0″ on kernel command line. boo#966255 * systemd 228 introduced cgroup Task accounting to limit fork bombs and the like. it seems a user logging in on the terminal is more limited than a GUI user, which could result in the system running out of resources to fork more tasks. boo#966878 And, last but not least, as promised, a word about why there have been no new snapshots released since 0209 (released on Feb 12): openSUSE Tumbleweed relies heavily on openQA and the test results we receive from the automated tests. For not yet clear reasons, the most powerful worker we have (had) on openQA.opensuse.org decided to quit its job (it did not mention the wish to find a new career path). This left openQA running with only 2 workers left instead of the usual 10. The remaining work force is largely overloaded and can’t cope with the workload. There were only two options left for us: release without testing or hold back the snapshots. As evidence shows, we opted to hold back the snapshots. Various solutions are currently evaluated to get new worker power on openQA, which includes borrowing machines from other SUSE owned instances. Have a great weekend ahead, Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Dear Tumbleweed users and hackers,
On 19 February 2016 at 15:54, Dominique Leuenberger / DimStar
Dear Tumbleweed users and hackers,
And, last but not least, as promised, a word about why there have been no new snapshots released since 0209 (released on Feb 12): openSUSE Tumbleweed relies heavily on openQA and the test results we receive from the automated tests. For not yet clear reasons, the most powerful worker we have (had) on openQA.opensuse.org decided to quit its job (it did not mention the wish to find a new career path). This left openQA running with only 2 workers left instead of the usual 10. The remaining work force is largely overloaded and can’t cope with the workload. There were only two options left for us: release without testing or hold back the snapshots. As evidence shows, we opted to hold back the snapshots. Various solutions are currently evaluated to get new worker power on openQA, which includes borrowing machines from other SUSE owned instances.
Have a great weekend ahead, Dominique
Since Dominique wrote this the SUSE Infra Team, Ludwig, and I have managed to migrate one of SUSE's internal SLE openQA workers to the openSUSE infra Thanks to all involved We now have a super-fast beast of a machine that is significantly faster than the old hardware running through the openqa.opensuse.org testing backlog, and I expect Tumbleweed to really enjoy and make the most out of this new hardware in the coming weeks. This will then be supplemented by even more new hardware sponsored by SUSE, which we are expecting will turn up soon (we ordered it some time ago, before the current problems with the current workers arose) Happy Hacking, Richard -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 19.02.16 schrieb Richard Brown:
This will then be supplemented by even more new hardware sponsored by SUSE, which we are expecting will turn up soon (we ordered it some time ago, before the current problems with the current workers arose)
I'd like to say a big thank you to SUSE. I think it is well deserved. Johannes -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 Comment: Using GnuPG with SeaMonkey - http://www.enigmail.net/ iEYEARECAAYFAlbHVLIACgkQzi3gQ/xETbKYYgCfaX23OEwKBOTFok9B4y7c3GRb 2E8An0UjO9bqqDidD/MTyti81cp2KTZM =0hFg -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Le 19/02/2016 18:32, Richard Brown a écrit :
time ago, before the current problems with the current workers arose)
you made a great job :-) I just have a clarification to ask. On your posts the word "worker" is a bit misleading, sometime it seems to be a computer and sometime a people. was it the case than an employee quit without warning and a computer break at the same time? bad :-( jdd -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Friday 19 February 2016 20:18:15 jdd wrote:
Le 19/02/2016 18:32, Richard Brown a écrit :
time ago, before the current problems with the current workers arose)
you made a great job :-)
I just have a clarification to ask. On your posts the word "worker" is a bit misleading, sometime it seems to be a computer and sometime a people.
was it the case than an employee quit without warning and a computer break at the same time? bad :-(
funny :-) I am pretty sure Richard is always talking about "worker" in the sense of "openQA-worker" as shown here: https://github.com/os-autoinst/openQA/blob/master/docs/GettingStarted.asciid... What Richard and Dominique *did* was to personify our "workers" or the machines they run on, not relating to employees. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Le 19/02/2016 20:22, Oliver Kurz a écrit :
On Friday 19 February 2016 20:18:15 jdd wrote:
Le 19/02/2016 18:32, Richard Brown a écrit :
time ago, before the current problems with the current workers arose)
you made a great job :-)
I just have a clarification to ask. On your posts the word "worker" is a bit misleading, sometime it seems to be a computer and sometime a people.
was it the case than an employee quit without warning and a computer break at the same time? bad :-(
funny :-) I am pretty sure Richard is always talking about "worker" in the sense of "openQA-worker" as shown here: https://github.com/os-autoinst/openQA/blob/master/docs/GettingStarted.asciid...
What Richard and Dominique *did* was to personify our "workers" or the machines they run on, not relating to employees.
may be, but what mean the fact than opensuse have only 2 workers when it should have 10? thanks jdd -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 19.02.16 schrieb jdd:
may be, but what mean the fact than opensuse have only 2 workers when it should have 10?
That the machine that was running the other 8 workers (VMs probably) died and thus only two workers (aka machines) are running at the moment (or rather as long, as the new machine is not running). Johannes -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 Comment: Using GnuPG with SeaMonkey - http://www.enigmail.net/ iEYEARECAAYFAlbHbuUACgkQzi3gQ/xETbKs1ACfaobeQ1plgjhyETBaqaOFpn58 +R4AnjdoQMnsKlXlpI3g/J0URyG/zIV5 =w2G5 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 19 February 2016 at 20:28, jdd
may be, but what mean the fact than opensuse have only 2 workers when it should have 10?
thanks jdd
openqa.opensuse.org used to have two physical servers, running a total of 10 openQA worker services each openQA "worker" is a service which runs a single openQA VM (or controls a single physical machine, or mainframe instance, or..now it's getting confusing, lets just keep it simple and talk about VMs) One machine ran 8 workers The other machine ran 2 workers The machine with 8 workers is misbehaving/broken 2 workers is not enough to test Tumbleweed (by the time it's finished testing, there are new things to test, so we never actually can release) The new servers (on order) can each run *at least* 16 workers (I suspect it can run more, but I designed them to run 16). And each work is also approximately twice as fast as any of the old workers. As of this afternoon, openqa.opensuse.org is using one machine, of a similar design, originally used by SUSE for the internal SLE openqa system, so we can now have 18 workers (16, plus the 2 old ones) When the new hardware arrives, openqa.opensuse.org will soon have 32 workers, split across two machines, so we won't be as vulnerable if one machine breaks again. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Le 19/02/2016 20:45, Richard Brown a écrit :
each openQA "worker" is a service which runs a single openQA VM (or controls a single physical machine, or mainframe instance, or..now it's getting confusing, lets just keep it simple and talk about VMs)
ok, got it :-) thanks
One machine ran 8 workers
The other machine ran 2 workers
The machine with 8 workers is misbehaving/broken
Murphy's law :-( but I prefer to see a hardware failure than a people failing :-)
When the new hardware arrives, openqa.opensuse.org will soon have 32 workers, split across two machines, so we won't be as vulnerable if one machine breaks again.
and much more power! thanks! jdd -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02/19/2016 09:54 AM, Dominique Leuenberger / DimStar wrote:
Dear Tumbleweed users and hackers,
Again I cover two weeks, with the 2nd one being more of troubles than releases. The snapshots covered this week are 0205, 0207, 0208, 0209 (week 6) and 0212 (week 7). I’ll try to explain at the end what was holding us back in week 7.
Noteworthy updates as part of the last snapshots:
* AppStream-glib update. Most notably: the icons are no longer duplicated across multiple repositories metadata * GCC 5.3.1 is now the default compiler * Mozilla Firefox 44 * cmake 3.4.3 – most importantly, ‘modules’ no longer get a soname during linking. Some packages might fail rpmlint now in case the shared library packaging policy was not applied * GStreamer 1.6.3 – Many bugfixes * Plasma 5.5.4 – you know you want it!
What is brewing in the next snapshot (0217+)
* Firewalld – available for testers. A future version should bring migration assistance from SUSEFirewall2 to firewalld * plymouth gained a new plugin to show simple texts (used to indicate which partition to unlock)
Have a great weekend ahead, Dominique
Any chance that Firewalld be written in Ruby so that it looks nice as the existing module SUSEFirewall2 in Yast? Cheers! Roman = -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02/19/2016 09:53 PM, Roman Bysh wrote:
Any chance that Firewalld be written in Ruby so that it looks nice as the existing module SUSEFirewall2 in Yast?
Cheers!
Roman
=
Hi Roman, I am not sure what you mean 'write firewalld in ruby'. I don't suppose you are suggesting to fork it and re-write to ruby do you? :) firewalld comes with a nice gtk3 UI interface. Look for the firewall-config package. My current plan is to extend the existing yast2-firewall module to use that (upstream) UI instead of writing a new one. It's still early days though. Give firewall-config a go if you haven't already. -- markos -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Mon, Feb 22, 2016 at 08:11:55AM +0000, Markos Chandras wrote:
On 02/19/2016 09:53 PM, Roman Bysh wrote:
Any chance that Firewalld be written in Ruby so that it looks nice as the existing module SUSEFirewall2 in Yast?
Cheers!
Roman
=
Hi Roman,
I am not sure what you mean 'write firewalld in ruby'. I don't suppose you are suggesting to fork it and re-write to ruby do you? :)
firewalld comes with a nice gtk3 UI interface. Look for the firewall-config package. My current plan is to extend the existing yast2-firewall module to use that (upstream) UI instead of writing a new one. It's still early days though. Give firewall-config a go if you haven't already.
SUSEfirewall2 is a shell script. yast2 firewall as yast2 module is in ruby. So I think the question might point to a yast2 module for firewalld ;) Ciao, Marcus -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02/22/2016 08:27 AM, Marcus Meissner wrote:
On Mon, Feb 22, 2016 at 08:11:55AM +0000, Markos Chandras wrote:
On 02/19/2016 09:53 PM, Roman Bysh wrote:
Any chance that Firewalld be written in Ruby so that it looks nice as the existing module SUSEFirewall2 in Yast?
Cheers!
Roman
=
Hi Roman,
I am not sure what you mean 'write firewalld in ruby'. I don't suppose you are suggesting to fork it and re-write to ruby do you? :)
firewalld comes with a nice gtk3 UI interface. Look for the firewall-config package. My current plan is to extend the existing yast2-firewall module to use that (upstream) UI instead of writing a new one. It's still early days though. Give firewall-config a go if you haven't already.
SUSEfirewall2 is a shell script.
yast2 firewall as yast2 module is in ruby.
So I think the question might point to a yast2 module for firewalld ;)
Ciao, Marcus
Yes. That is what I was trying to say. -- Cheers! Roman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 19/02/16 14:54, Dominique Leuenberger / DimStar wrote:
What is brewing in the next snapshot (0217+)
* Firewalld â available for testers. A future version should bring migration assistance from SUSEFirewall2 to firewalld
I have uploaded an initial version of this script to https://github.com/openSUSE/susefirewall2-to-firewalld As usual, use './susefirewall2-to-firewalld -h' as a first step ;) Remember, this script will probably work for simple SF2 configurations but may break for complex ones. It's goal is to provide an _initial_ firewalld configuration close to what you had with SF2 so you should double check the end result yourself. Feel free to give it a go and report bugs either on github or bugzilla.opensuse.org. Unless a major blocking bug is found, I plan to tag the first release later this month and package it in OBS. -- markos -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (8)
-
Dominique Leuenberger / DimStar -
jdd -
Johannes Kastl -
Marcus Meissner -
Markos Chandras -
Oliver Kurz -
Richard Brown -
Roman Bysh