Issue #19162 has been updated by Curtis Ruck.
i've found this can be done by chaining stages together, with each stage having
a refreshonly=true exec to force reboot. It shows failed a few times until
it gets through all the stages though.
Feature
Issue #19379 has been reported by Chris Ritson.
Bug #19379: exec: logoutput=on_error always logs
https://projects.puppetlabs.com/issues/19379
Author: Chris Ritson
Status: Unreviewed
Priority: Normal
Assignee:
Category:
Target version:
Affected Puppet
Issue #7680 has been updated by Chris Boot.
Hi,
I'm struggling to understand the updates from Adrien to this bug; I can confirm
again that applying the patch to Puppet *clients* running version 2.7.18-2 (as
packaged by Debian) resolves the problem for us and so far we have seen no
adverse
Issue #12179 has been updated by Henrik Lindberg.
Suggest using a repository manager that includes a feature to define
(constrained) views of the repository. Manage the constrained sets with the
repository manager.
(Take a look at something like Sontatype's Nexus)
Issue #14784 has been updated by Joachim de Groot.
Title and description are misleading as the problem in this case is the epoch
of the package.
The JDK rpm package has the epoch of **2000**, which seems to cause these
issues:
/bin/rpm -q jdk --nosignature --nodigest --qf '%{NAME}
Issue #12556 has been updated by Andre Nathan.
Any work being done on this? I believe the facts should contain a list of
addresses, since multiple addresses in a single interface are supported by IPv6.
Bug #12556: ipv6 address facts
Issue #19376 has been updated by Curtis Ruck.
This appears to be a relative of #9622 from PE.
Bug #19376: user/group Autorequire with resources{'group':purge=true} and
resources{'user':purge=true}
Issue #2823 has been updated by Alexander Komarov.
This problem still occurs in puppet 2.7.20 (not sure where to file this), but
more importantly, it occurs when the catalog in question is excluded via tags,
causing a major issue for selective deployments:
This works (when not using tags),
Issue #5517 has been updated by John Bollinger.
Luke Kanies wrote:
John Bollinger wrote:
Do you mean *any* class that hasn't yet been evaluated, or do you mean the
class to which the variable belongs? If the latter, then you are creating
a new failure scenario.
I just mean
Issue #19188 has been updated by micah -.
So you would like to have something like a global default?
Yeah, that is the point, a global default so you don't need to define a default
for each key.
Why don’t you define a default yourself directly in the code?
I do that, it works, but if I
Issue #2823 has been updated by Alexander Komarov.
UPDATE: I found one workaround (use a failover directory):
source = ['/tmp/nonexist1', '/some/failover/directory/that/exists']
Now eval_generate() will not fail, but I still don't understand why puppet
wants to traverse the whole source
Issue #19385 has been reported by Yannig rousseau.
Bug #19385: Puppetdb installation via module
https://projects.puppetlabs.com/issues/19385
Author: Yannig rousseau
Status: Unreviewed
Priority: Normal
Assignee:
Category:
Target version:
Affected
Issue #19385 has been updated by Ken Barber.
Status changed from Unreviewed to Rejected
Hiya Yannig, this problem should be logged here:
https://github.com/puppetlabs/puppet-postgresql/issues?state=open
Thanks!
Bug #19385: Puppetdb installation via
Issue #15561 has been updated by Jason Hancock.
You can also run into this bug when using the http report processor submitting
to an https reporturl where the certificate has a '/' in it.
Bug #15561: Fix for CVE-2012-3867 is too restrictive
Issue #19390 has been reported by Brian Rak.
Bug #19390: Puppet agent getting stuck
https://projects.puppetlabs.com/issues/19390
Author: Brian Rak
Status: Unreviewed
Priority: Normal
Assignee:
Category:
Target version:
Affected Puppet version: 3.1.0
Issue #19390 has been updated by Brian Rak.
Realized I forgot this, but these were run while stuck:
pre# puppet config print | grep agent_catalog_run_lockfile
agent_catalog_run_lockfile = /var/lib/puppet/state/agent_catalog_run.lock
# ls -l /var/lib/puppet/state/agent_catalog_run.lock
ls:
Issue #19390 has been updated by Brian Rak.
Kernel version: 2.6.18-348.el5 x86_64
Bug #19390: Puppet agent getting stuck
https://projects.puppetlabs.com/issues/19390#change-83674
Author: Brian Rak
Status: Unreviewed
Priority: Normal
Assignee:
Category:
Issue #17879 has been updated by Yuri Arabadji.
Status changed from Closed to Re-opened
Alrighty. Made a pull request.
a
href=https://github.com/puppetlabs/puppet/pull/1490;https://github.com/puppetlabs/puppet/pull/1490/a
Bug #17879: extract cert name
Issue #15561 has been updated by Yuri Arabadji.
okay, what if you try a
href=https://github.com/puppetlabs/puppet/pull/1490;https://github.com/puppetlabs/puppet/pull/1490/a
? tnx.
Bug #15561: Fix for CVE-2012-3867 is too restrictive
Issue #19395 has been reported by Aaron Grewell.
Feature #19395: Detect client-server version mismatch, warn and optionally
disallow
https://projects.puppetlabs.com/issues/19395
Author: Aaron Grewell
Status: Unreviewed
Priority: Normal
Assignee:
Issue #19396 has been reported by David K.
Bug #19396: Wrong environment being saved in yaml/database when using
storeconfigs
https://projects.puppetlabs.com/issues/19396
Author: David K
Status: Closed
Priority: Normal
Assignee:
Category:
Target
Issue #19396 has been updated by David K.
Status changed from Closed to Re-opened
Not sure why it got closed
Bug #19396: Wrong environment being saved in yaml/database when using
storeconfigs
https://projects.puppetlabs.com/issues/19396#change-83682
Issue #19396 has been updated by eric sorenson.
Status changed from Re-opened to Investigating
Assignee set to eric sorenson
Going to try to reproduce this under puppetdb.
Bug #19396: Wrong environment being saved in yaml/database when using
Issue #19396 has been updated by eric sorenson.
This doesn't reproduce for me w/puppetdb so unless I've missed something it's
specific to activerecord stored configs and therefore not very likely to be
fixed.
Bug #19396: Wrong environment being saved
Issue #17879 has been updated by Jeff McCune.
Status changed from Re-opened to Duplicate
Unfortunately the pull request in
https://github.com/puppetlabs/puppet/pull/1490 does not sufficiently address
the root cause of this problem.
Here's the comment I posted on Github:
Thanks for this
Issue #15561 has been updated by Jeff McCune.
Yuri Arabadji wrote:
okay, what if you try a
href=https://github.com/puppetlabs/puppet/pull/1490;https://github.com/puppetlabs/puppet/pull/1490/a
? tnx.
Unfortunately this change does not sufficiently address the root cause of the
problem,
Issue #14093 has been updated by Jeff McCune.
Category set to templates
Status changed from In Topic Branch Pending Review to Merged - Pending Release
Target version set to 3.2.0
Merged into master as 432850fb74813eded3036f861e05d9266289c16c.
This should be released in 3.2.0.
Thanks again for
Issue #19247 has been updated by Adrien Thebo.
I was able to track down the original backstory behind this change.
- - -
The change to console logging conventions was a part of a couple of overlapping
discussions.
Firstly, the work being done on the module tool had need of undecorated
Issue #7243 has been updated by Adrien Thebo.
Hi Patrick,
Would you be able to file a pull request against the Github repo
(https://github.com/puppetlabs/puppet) ? This well help us start reviewing the
code on a more formal level. You might want to review
Issue #5517 has been updated by Luke Kanies.
You're right that it's basically impossible to escape pathological cases.
I think we can get fantastically close by switching to lazy on-demand
evaluation and then at least throwing a warning if someone tries to override a
class that's already
Issue #17030 has been updated by Adrien Thebo.
I spoke with Jeff about this and the preferred resolution is to try pip if it's
available and then fall back to pip-python on Redhat. On other platforms we
just just probe for pip.
In addition it might be worthwhile to check the pip version to
Issue #19315 has been updated by James Shubin.
You do have some interesting points. Hopefully I'll be able to address them
adequately:
I guess the reason why I think it makes sense is that I see $name as a
variable, and logically it makes sense to allow any first class type. (Why
should a
32 matches
Mail list logo