https://bugzilla.redhat.com/show_bug.cgi?id=1304882
awill...@redhat.com changed:
What|Removed |Added
Status|ASSIGNED|CLOSED
Resolution|---
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #56 from Jon Ciesla ---
Package request has been approved:
https://admin.fedoraproject.org/pkgdb/package/openqa
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #55 from awill...@redhat.com ---
Thanks very much for the review, both of you!
I want to write an SELinux policy but I don't think I'll block the package for
that.
--
You are receiving this mail because:
You are on the CC list f
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
John Dulaney changed:
What|Removed |Added
Flags|fedora-review? |fedora-review+
--- Comment #54 from Jo
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #53 from Zbigniew Jędrzejewski-Szmek ---
I think the package looks good.
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #52 from awill...@redhat.com ---
John: Zbigniew: where do we stand with this? Do you have outstanding concerns
with 4.3-12? For the record, https://openqa.stg.fedoraproject.org has been
using 4.3-12 for the last few days and seems
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #51 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-12.test.3.fc23.src.rpm for f23 completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12957773
--
You are receiving this mail because:
You are
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #50 from awill...@redhat.com ---
The .test.X scratch builds are me testing a complex patch I'm working on
upstream, not directly related to package review.
--
You are receiving this mail because:
You are on the CC list for the bu
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #49 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-12.test.1.fc23.src.rpm for f23 completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12957108
--
You are receiving this mail because:
You are
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #48 from awill...@redhat.com ---
A small note on the trigger stuff: it's not clean on upgrade from the old
system - the bit of RPM that says "remove this directory that's no longer
packaged" runs after the %triggerin script, so whe
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #47 from awill...@redhat.com ---
On the UID/GID issue: upstream merged my patch to stop initdb attempting to
switch uids/gids twice, which I think is a decent fix. It means the db file
winds up owned by geekotest.geekotest (not gee
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #46 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-12.fc23.src.rpm for f23 completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12956321
--
You are receiving this mail because:
You are on the
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #45 from Zbigniew Jędrzejewski-Szmek ---
I think the asset generation script should remove
/usr/share/openqa/public/sass/.sass-cache/ after it's done. That directory has
mode 0700 which is annoying.
--
You are receiving this mail
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #44 from Zbigniew Jędrzejewski-Szmek ---
(In reply to awill...@redhat.com from comment #36)
> Another fix, I guess, would be to have /var/lib/openqa/db owned by
> geekotest.geekotest , which would happen to work OK with the code as
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #43 from John Dulaney ---
Just out of curiosity, what happens if you leave the db directory owned by
geekotest.root and then set facls to allow the geekotest user write access?
--
You are receiving this mail because:
You are on t
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #42 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-12.fc23.src.rpm for f23 completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12955213
--
You are receiving this mail because:
You are on the
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #41 from awill...@redhat.com ---
Well, here's some crap that might be better, I guess?
https://www.happyassassin.net/reviews/openqa/openqa.spec
https://www.happyassassin.net/reviews/openqa/openqa-4.3-11.fc23.src.rpm
https://www.ha
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #40 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-11.fc23.src.rpm for f23 completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12954670
--
You are receiving this mail because:
You are on the
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #39 from awill...@redhat.com ---
Note while I work on this stuff: the trigger approach has one drawback, which
is that it requires a regular dependency on rubygem(sass) - as the generation
is now done on the target system and not i
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #38 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-11.fc23.src.rpm for f23 completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12953562
--
You are receiving this mail because:
You are on the
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #37 from awill...@redhat.com ---
There's an irony here: if we change initdb to accommodate database.ini *not*
being owned by geekotest, we might actually *create* an attack vector in the
case where it *is* owned by geekotest. The p
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #36 from awill...@redhat.com ---
OK, so I worked this out, but it's a bit...icky. It's indeed caused by
database.ini being not owned by geekotest.
initdb has a `--user` argument used to run as a particular user. It switches
UID to
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #35 from awill...@redhat.com ---
If I modify the test script to throw an open() in there:
#!/bin/perl
use Config::IniFiles;
use POSIX qw/setuid setgid/;
setuid 995;
setgid 994;
open( my $fh, '<', '/etc/openqa/database.ini' );
w
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #34 from awill...@redhat.com ---
So, uh, yeah, this is the problem, but it's rather odd:
http://fpaste.org/321675/52705111/
why does that work as geekotest but not as root?! Some subtlety of setuid /
setgid that I don't understan
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #33 from awill...@redhat.com ---
Yeah, that's what it is. This fails when database.ini is root.geekotest 0640:
/usr/share/openqa/script/initdb --user geekotest --init_database
I think it's because of the way initdb drops privileg
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #32 from awill...@redhat.com ---
Those errors look like a case where OPENQA_CONFIG isn't set (if it is set,
Schema.pm doesn't look for ../etc/openqa/database.ini). I can't see why that
would happen in your script.
It's *possible*
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #31 from awill...@redhat.com ---
"/var/lib/openqa/share/factory/iso is not %ghosted, on purpose?" Yeah, I kinda
go back and forward on that one. It's somewhat different because the upstream
Makefile actually installs it and it's mo
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #30 from Zbigniew Jędrzejewski-Szmek ---
Created attachment 1123306
--> https://bugzilla.redhat.com/attachment.cgi?id=1123306&action=edit
patch to use triggers to generate assets
/var/lib/openqa/share/factory/iso is not %ghosted
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #29 from awill...@redhat.com ---
Forgot to mention - openqa logging is pretty light by default (info level).
debug level is very chatty, but it's not the default.
--
You are receiving this mail because:
You are on the CC list for
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #28 from awill...@redhat.com ---
So we can achieve fairly simple logging to the journal simply by setting the
log 'file' setting to nothing at all in config:
file =
or making it undefined it in the code block that defines the 'de
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #27 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-10.fc23.src.rpm for f23 completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12934037
--
You are receiving this mail because:
You are on the
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #26 from Zbigniew Jędrzejewski-Szmek ---
(In reply to Ludwig Nussel from comment #22)
> If you prefer logging to the journal ie stdout feel free let your openQA
> package do that instead.
That would seem like the best option, unle
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #25 from awill...@redhat.com ---
The apache case is a bit different because IIRC Apache starts as root and
creates missing log files as root before it drops privs; /var/log/httpd is
owned by root. The reason not to create /var/log/
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #24 from John Dulaney ---
I note that the httpd rpm install /var/log/httpd (the folder), but no log files
are populated until the httpd server is started. At that point, it is
populated. However, uninstalling httpd leaves the log
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #23 from Ludwig Nussel ---
(In reply to Zbigniew Jędrzejewski-Szmek from comment #13)
> > "Strictly speaking, the generator is wrong, because generators cannot rely
> > on /var being mounted. It will not operate correctly if someon
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
Ludwig Nussel changed:
What|Removed |Added
CC||ludwig.nus...@suse.de
--- Comment #22
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #21 from John Dulaney ---
I think I prefer 4 over 5; it can make it easier to scrape logs as necessary.
For 3, how often do you plan to uninstall openqa? It's sort of like setting up
a web server and then uninstalling apache. 2
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #20 from awill...@redhat.com ---
so sadly, just ditching the log file stuff doesn't work, because geekotest of
course can't create /var/log/openqa (as it doesn't have the necessary
permissions on /var/log). d'oh. So effectively we
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #19 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-9.fc23.src.rpm for f23 completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12916576
--
You are receiving this mail because:
You are on the
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #18 from Upstream Release Monitoring
---
adamwill's scratch build of openqa-4.3-9.fc23.src.rpm for f23 failed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12916515
--
You are receiving this mail because:
You are on the CC
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #17 from awill...@redhat.com ---
So I've put up a 4.3-9 with several changes:
https://www.happyassassin.net/reviews/openqa/openqa.spec
https://www.happyassassin.net/reviews/openqa/openqa-4.3-9.fc23.src.rpm
I haven't actually *tes
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
Zbigniew Jędrzejewski-Szmek changed:
What|Removed |Added
Assignee|nob...@fedoraproject.org|jdula...@fedoraproject.org
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #15 from awill...@redhat.com ---
"%config(noreplace) %attr(-,geekotest,root) %{_sysconfdir}/openqa/openqa.ini
%config(noreplace) %attr(-,geekotest,root) %{_sysconfdir}/openqa/database.ini
Why are those owned by the user? Can they b
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #14 from awill...@redhat.com ---
So the %systemd_requires thing is only in a draft page so far as I can see -
https://fedoraproject.org/wiki/User:Zbyszek/SystemdLinksDraft - but I'm fine
with expanding it.
On the generator: I thin
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #13 from Zbigniew Jędrzejewski-Szmek ---
(In reply to awill...@redhat.com from comment #12)
> "%{?systemd_requires} is forbidden by the guidelines. I don't think we gain
> anything by that rule, but it's on the books."
>
> Thanks,
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #12 from awill...@redhat.com ---
"%{?systemd_requires} is forbidden by the guidelines. I don't think we gain
anything by that rule, but it's on the books."
Thanks, I'll...er...do something about that?
"What about parallel build?"
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #9 from Neal Gompa ---
Adam, the URLs all return HTTP 403 Forbidden.
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product and component
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #10 from awill...@redhat.com ---
goddamnit, selinux...OK, fixed (it was only the spec actually).
--
You are receiving this mail because:
You are on the CC list for the bug.
You are always notified about changes to this product an
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #8 from awill...@redhat.com ---
As for system-generators - there's a file in there,
/usr/lib/systemd/system-generators/systemd-openqa-generator . The install
process installs it. I'm not sure what it's for, but...something? :)
--
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #7 from awill...@redhat.com ---
I'm not actually sure what happens if the assets become invalidated and the
re-generation fails - it may just use the underlying files and keep working
(with a logged warning or so), or it may result
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #11 from Zbigniew Jędrzejewski-Szmek ---
%{?systemd_requires} is forbidden by the guidelines. I don't think we gain
anything by that rule, but it's on the books.
What about parallel build?
I think user/group creation scriptlets s
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #6 from John Dulaney ---
What is %{_prefix}/lib/systemd/system-generators used for? I see it defined,
and that's it.
Also, I note that if I get reverse-depcheck finished, new builds of any
packages required by %requires_eq will w
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #5 from Upstream Release Monitoring
---
jdulaney's scratch build of openqa-4.3-7.fc23.src.rpm for rawhide completed
http://koji.fedoraproject.org/koji/taskinfo?taskID=12894011
--
You are receiving this mail because:
You are on t
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #4 from awill...@redhat.com ---
Here's 4.3-7, with all of Neal's comments addressed, and a couple of other
changes:
- package review improvements:
- * no need for worker to Requires(post) os-autoinst
- * explain why tests are curr
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
John Dulaney changed:
What|Removed |Added
Status|NEW |ASSIGNED
CC|
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
--- Comment #3 from awill...@redhat.com ---
1. Good point on the Requires(post), I'm not entirely sure why it's there; I
don't see why os-autoinst would be required for %post, unless it's somehow
needed for database creation (I don't see why).
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
Neal Gompa changed:
What|Removed |Added
CC||ngomp...@gmail.com
--- Comment #2 from N
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
awill...@redhat.com changed:
What|Removed |Added
See Also||https://bugzilla.redhat.com
https://bugzilla.redhat.com/show_bug.cgi?id=1304882
awill...@redhat.com changed:
What|Removed |Added
CC||mgr...@redhat.com,
59 matches
Mail list logo
