All but one of these tests is fixed by:
* Stubbing Puppet.settings.use
* Adding /usr/sbin to PATH
The only other one was the package integration test,
which stupidly assumed a default was specified in the test.
The fix here is twofold: Remove that assumption (the test is
now 'pending' if no defau
These tests once again largely were caused
by /usr/sbin not being in the path and by
~ not being writable.
The only tests still failing are Rails tests,
and my guess is that they're all failing because
of the recent work by Brice. They should probably
just be removed.
Signed-off-by: Luke Kanies
This includes my previous patch plus fixes for test/.
spec/integration/checksum.rb |1 +
spec/integration/reports.rb|4
spec/integration/transaction/report.rb |8 +++-
spec/integration/type/package.rb |7 +--
spec/unit/indirector/ssl_rsa/file.r
All but one of these tests is fixed by:
* Stubbing Puppet.settings.use
* Adding /usr/sbin to PATH
The only other one was the package integration test,
which stupidly assumed a default was specified in the test.
The fix here is twofold: Remove that assumption (the test is
now 'pending' if no defau
On Feb 27, 2009, at 10:29 PM, Nigel Kersten wrote:
>
> On Fri, Feb 27, 2009 at 4:45 PM, Luke Kanies wrote:
>
>> The big reason I thought about using parameters was just that we
>> might
>> need to version the overall API (the split between environment,
>> resource type, and resource key) separ
On Fri, Feb 27, 2009 at 4:45 PM, Luke Kanies wrote:
> The big reason I thought about using parameters was just that we might
> need to version the overall API (the split between environment,
> resource type, and resource key) separately from the resource type API
> (any formatting within the key
On Fri, Feb 27, 2009 at 06:45:12PM -0600, Luke Kanies wrote:
>
> On Feb 27, 2009, at 1:07 PM, Eric Gerlach wrote:
>
> >
> > On Fri, Feb 27, 2009 at 12:47:36PM -0600, Luke Kanies wrote:
> >> Your points are definitely valid. This is the closest (and best)
> >> discussion I've found:
> >>
> >> ht
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi all
What's everyone's thoughts/feelings/etc on ticket bounties?
I've been thinking about the many issues assigned to community currently
and wondering if a bounty system might not help here (mightn't either
I'll admit).
It'd work something like:
On Feb 27, 2009, at 1:07 PM, Eric Gerlach wrote:
>
> On Fri, Feb 27, 2009 at 12:47:36PM -0600, Luke Kanies wrote:
>> Your points are definitely valid. This is the closest (and best)
>> discussion I've found:
>>
>> http://stackoverflow.com/questions/389169/best-practices-for-api-versioning
>
> Ac
I can't reproduce this. You still have this issue with current HEAD?
On Feb 21, 2009, at 4:41 PM, Brice Figureau wrote:
>
> Hi,
>
> I was trying to use current master HEAD with this simple manifest:
> node "macbook.local" {
> #notify { "yep": }
> file {
> "/tmp/c": content => "p
Signed-off-by: Luke Kanies
---
lib/puppet/defaults.rb |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/puppet/defaults.rb b/lib/puppet/defaults.rb
index ef17001..0646943 100644
--- a/lib/puppet/defaults.rb
+++ b/lib/puppet/defaults.rb
@@ -384,8 +384,6 @@ module Pup
The first patch is just a fix to the puppetd setup.
lib/puppet/defaults.rb |4 +-
lib/puppet/ssl/host.rb | 15 -
spec/unit/ssl/host.rb | 78 +++-
3 files changed, 93 insertions(+), 4 deletions(-)
--~--~-~--~~~
The problem was that the server had a certificate
for the client. Initially the client just didn't have a
key, because it assumed that if it had a certificate then
it had a key. Upon fixing it to create the key, the key then
did not match the found certificate.
This commit fixes both of those:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Brice Figureau wrote:
> This script fixes the most common issues with inconsistent
> storeconfigs database (including duplicate resources record,
> duplicate param_values records, dangling records...).
>
Pushed in commit ec56ddf8f5b58f16d00670553468
On Fri, Feb 27, 2009 at 3:04 PM, Luke Kanies wrote:
>>> Got a reasonable alternate executable proposal? Something we could
>>> reasonably use to, at least, merge puppetca and puppetclean, and
>>> probably merge filebucket into it, too. I could see adding ralsh to
>>> it, also.
>>> puppetcli?
>>
On Feb 27, 2009, at 4:55 PM, James Turnbull wrote:
>
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Luke Kanies wrote:
>> It probably makes sense to have a plan first, then worry about
>> implementing it.
>>
>> I agree that it would be confusing to have the 'puppet' executable do
>> all of
I think we should follow guidelines where we can, but I don't think we
should get crazy on it. The whitespace changes here don't hurt the
readability of the patch, so I'm fine either way.
I think separation makes sense when the main changes are swamped by
the secondary (whitespace, formatt
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Luke Kanies wrote:
> It probably makes sense to have a plan first, then worry about
> implementing it.
>
> I agree that it would be confusing to have the 'puppet' executable do
> all of this.
>
> Got a reasonable alternate executable proposal?
On Feb 27, 2009, at 4:21 PM, Brice Figureau wrote:
>> [...]
>> We've got this pretty consistent model across multiple of our
>> executables: A given kind of functionality (filebucket, ca, node, and
>> maybe more), each with their own commands.
>>
>> What do you think?
>
> Hm, git unlike puppet (th
+1
On Feb 27, 2009, at 4:05 PM, Brice Figureau wrote:
>
> This script fixes the most common issues with inconsistent
> storeconfigs database (including duplicate resources record,
> duplicate param_values records, dangling records...).
>
> Usage:
> stop all puppetmasters
> backup your database!
On Feb 27, 2009, at 3:44 PM, Gary Law wrote:
>
>
> Would everyone be using pkgutil, such that it's worth making that
> dependency?
>
> It's perfectly possible to install the opencsw/blastwave
> repositories and not have pkgutil (or pkg-get). What will happen is
> if they use pkg-get to get pup
On 27/02/09 19:11, Luke Kanies wrote:
> In thinking about this, doesn't it make sense to have more than just
> cleaning as a node-level operation?
>
> That is, maybe have a 'puppetnode' executable or something silly, and
> it could accept a '--clean' option, but could also list, get status,
This script fixes the most common issues with inconsistent
storeconfigs database (including duplicate resources record,
duplicate param_values records, dangling records...).
Usage:
stop all puppetmasters
backup your database!
% cat ext/dbfix.sql | mysql puppet
relaunch all puppetmasters
Signed-o
On 27/02/09 21:46, Luke Kanies wrote:
> Looks good, but I'd like to hear testing results from someone with
> possible high loads - isn't there a chance concurrency could be an
> issue with this?
Yes, I'd like to know the answer too, hence my warning.
I'm quite ignorant about mongrel (or webr
2009/2/27 Luke Kanies
>
> On Feb 27, 2009, at 1:39 PM, Gary Law wrote:
>
> > 2009/2/27 Luke Kanies
> > You keep switching between 'pkguti' and 'pkgutil' here -- is that
> > discrepancy intentional or accidental?
> >
> >
> > Hi
> >
> > Yeah, there was the use of pkg-get and pkgget as names in the
master does the right thing:
if time < 1
puts "Exiting; no certificate found and waitforcert is disabled"
exit(1)
end
On Fri, Feb 27, 2009 at 1:05 PM, Paul Lathrop wrote:
> On Fri, Feb 27, 2009 at 12:48 PM, Luke Kanies wrote:
>>
>> Note that the certhand
Okay, well I am just trying to follow the patch guidelines ;-) I've
pushed a patch w/o the whitespace changes to my github:
http://github.com/plathrop/puppet/commits/ticket/0.24.x/2031 commit
fcdd9e68cd06cd007e3fbc618a87b4d746035d13
Personally I like cleaning up whitespace as I go. Your call.
-
I'm fine removing it; just wanted to be clear it was a positive
change. :)
On Feb 27, 2009, at 3:08 PM, Paul Lathrop wrote:
>
> On Fri, Feb 27, 2009 at 1:07 PM, Luke Kanies wrote:
>>
>> What's up with all the whitespace changes in this patch?
>
> Oh dammit!
>
> I use emacs with delete-trailin
On Fri, Feb 27, 2009 at 1:07 PM, Luke Kanies wrote:
>
> What's up with all the whitespace changes in this patch?
Oh dammit!
I use emacs with delete-trailing-whitespace as a save hook.
I'll fix the patch (again).
Sorry.
--Paul
--~--~-~--~~~---~--~~
You receive
What's up with all the whitespace changes in this patch?
On Feb 27, 2009, at 3:03 PM, Paul Lathrop wrote:
>
> "You can turn off waiting for certificates by specifying a time of 0."
>
> Also add a test to ensure we catch any future regression of this
> behavior.
>
> Signed-off-by: Paul Lathrop
This version doesn't have the typo.
On Fri, Feb 27, 2009 at 1:03 PM, Paul Lathrop wrote:
>
> "You can turn off waiting for certificates by specifying a time of 0."
>
> Also add a test to ensure we catch any future regression of this behavior.
>
> Signed-off-by: Paul Lathrop
> ---
> lib/puppet/
On Fri, Feb 27, 2009 at 12:48 PM, Luke Kanies wrote:
>
> Note that the certhandler module is gone in master; most of the
> functionality has moved to ssl/host.rb.
I'll make sure the behavior is correct there and submit a separate
patch if appropriate.
--Paul
--~--~-~--~~---
"You can turn off waiting for certificates by specifying a time of 0."
Also add a test to ensure we catch any future regression of this behavior.
Signed-off-by: Paul Lathrop
---
lib/puppet/executables/client/certhandler.rb |7 ++-
spec/unit/executables/client/certhandler.rb | 19 +++
Note that the certhandler module is gone in master; most of the
functionality has moved to ssl/host.rb.
On Feb 27, 2009, at 2:33 PM, Paul Lathrop wrote:
>
> Also add a test to ensure we catch any future regression of this
> behavior.
>
> Signed-off-by: Paul Lathrop
> ---
> lib/puppet/execut
+1
On Feb 27, 2009, at 2:10 PM, Nigel Kersten wrote:
>
>
> Signed-off-by: Nigel Kersten
> ---
> conf/osx/preflight |2 +-
> 1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/conf/osx/preflight b/conf/osx/preflight
> index 9b2c07e..98251bf 100755
> --- a/conf/osx/preflight
> +
Looks good, but I'd like to hear testing results from someone with
possible high loads - isn't there a chance concurrency could be an
issue with this?
On Feb 27, 2009, at 2:04 PM, Brice Figureau wrote:
>
> Rationale: Rails likes to cache a lot of things to improve
> performance of OLTP appli
On Feb 27, 2009, at 1:39 PM, Gary Law wrote:
> 2009/2/27 Luke Kanies
> You keep switching between 'pkguti' and 'pkgutil' here -- is that
> discrepancy intentional or accidental?
>
>
> Hi
>
> Yeah, there was the use of pkg-get and pkgget as names in the
> original. pkgget looked like an alias f
Also add a test to ensure we catch any future regression of this behavior.
Signed-off-by: Paul Lathrop
---
lib/puppet/executables/client/certhandler.rb |7 ++-
spec/unit/executables/client/certhandler.rb | 13 +
2 files changed, 19 insertions(+), 1 deletions(-)
diff --gi
Signed-off-by: Nigel Kersten
---
conf/osx/preflight |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/conf/osx/preflight b/conf/osx/preflight
index 9b2c07e..98251bf 100755
--- a/conf/osx/preflight
+++ b/conf/osx/preflight
@@ -9,4 +9,4 @@
# when being installed to volume
Rationale: Rails likes to cache a lot of things to improve
performance of OLTP applications. Puppet is a long
lived server which doesn't really have an OLTP database workload.
The idea is to close the connection to the database after each
compilation of a node to flush all the active cache to rele
On Fri, Feb 27, 2009 at 10:56 AM, Luke Kanies wrote:
> Do you really care if waitforcert 0 is missing, since waitforcert 1 is
> essentially equivalent?
Good point. No I don't care.
> In looking at other timeout values (e.g., filetimeout) we usually use
> '0' to mean disabling it, so I think we
2009/2/27 Luke Kanies
> You keep switching between 'pkguti' and 'pkgutil' here -- is that
> discrepancy intentional or accidental?
Hi
Yeah, there was the use of pkg-get and pkgget as names in the original.
pkgget looked like an alias for pkg-get, but my ruby isn't good enough to be
sure, so I
On Fri, Feb 27, 2009 at 12:47:36PM -0600, Luke Kanies wrote:
> Your points are definitely valid. This is the closest (and best)
> discussion I've found:
>
> http://stackoverflow.com/questions/389169/best-practices-for-api-versioning
Actually, that's a really good treatment of the subject.
>
On Feb 26, 2009, at 4:48 PM, Gary Law wrote:
> Hi
> I've written a package provider. It's intended to replace or augment
> the pkg-get based blastwave provider (which doesn't work properly,
> because pkg-get doesn't work properly).
> It's intended to be a minimal change from what's in blastwave.r
On Feb 26, 2009, at 12:39 PM, Brice Figureau wrote:
>
> On 26/02/09 19:28, Paul Lathrop wrote:
>> So, I realized there are two reasonable ways to resolve
>> http://projects.reductivelabs.com/issues/2031
>>
>> 1) Change the documentation to match the actual behavior.
>>
>> 2) Re-implement the spec
On Feb 26, 2009, at 12:28 PM, Paul Lathrop wrote:
>
> So, I realized there are two reasonable ways to resolve
> http://projects.reductivelabs.com/issues/2031
>
> 1) Change the documentation to match the actual behavior.
>
> 2) Re-implement the special-case behavior of 'waitforcert 0'
>
> I'm lean
On Feb 26, 2009, at 10:58 AM, Eric Gerlach wrote:
>
> On Sun, Feb 22, 2009 at 12:02:08PM -0800, Luke Kanies wrote:
>>
>> On Feb 21, 2009, at 6:24 PM, Ohad Levy wrote:
>>
>>>
>>>
>>> On Sun, Feb 22, 2009 at 2:42 AM, Luke Kanies
>>> wrote:
>>>
>>>
>>> I could see some other version numbers makin
On Feb 23, 2009, at 1:10 PM, Nigel Kersten wrote:
>
> On Sun, Feb 22, 2009 at 12:02 PM, Luke Kanies
> wrote:
>>
>> On Feb 21, 2009, at 6:24 PM, Ohad Levy wrote:
>>
>>>
>>>
>>> On Sun, Feb 22, 2009 at 2:42 AM, Luke Kanies
>>> wrote:
>>>
>>>
>>> I could see some other version numbers making t
On Feb 21, 2009, at 8:53 AM, Brice Figureau wrote:
>
> Yet, only the store reports has an implementation.
>
> Signed-off-by: Brice Figureau
> ---
> lib/puppet/indirector/report/processor.rb | 10 ++-
> lib/puppet/reports/log.rb |3 ++
> lib/puppet/reports/rrdgraph.rb
+1
On Feb 21, 2009, at 8:53 AM, Brice Figureau wrote:
>
> we already offer read access for the yaml, this changeset adds
> the destroy handler which only removes the yaml file for
> a request. This can be used to remove cached entries.
>
> Signed-off-by: Brice Figureau
> ---
> lib/puppet/indire
+1
On Feb 21, 2009, at 8:53 AM, Brice Figureau wrote:
>
> The new name should be a symbol to be able to access the
> configuration parsed from the settings file.
>
> Signed-off-by: Brice Figureau
> ---
> lib/puppet/util/settings.rb |4 +++-
> spec/unit/util/settings.rb | 11 +++
>
In thinking about this, doesn't it make sense to have more than just
cleaning as a node-level operation?
That is, maybe have a 'puppetnode' executable or something silly, and
it could accept a '--clean' option, but could also list, get status,
etc.
We could start with what you have, but I'
On Feb 27, 2009, at 9:46 AM, James Turnbull wrote:
>
>
> Signed-off-by: James Turnbull
> ---
> CHANGELOG |3 +++
> lib/facter/util/uptime.rb |3 ++-
> 2 files changed, 5 insertions(+), 1 deletions(-)
>
> diff --git a/CHANGELOG b/CHANGELOG
> index fc0d4d7..ac03ca6 100644
> -
Signed-off-by: James Turnbull
---
CHANGELOG |3 +++
lib/facter/util/uptime.rb |3 ++-
2 files changed, 5 insertions(+), 1 deletions(-)
diff --git a/CHANGELOG b/CHANGELOG
index fc0d4d7..ac03ca6 100644
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -1,3 +1,6 @@
+1.5.x:
+ Fixed #203
On Feb 27, 2009, at 8:28 AM, Brice Figureau wrote:
>>
> Does that mean Luke's patch adds more scalability?
as far as memory any little thing can help, so yes.
>
>
>> - storedconfig is foobared and sucks memory like you can't believe.
>> Once I disabled the storedconfig memory did not grow like
On Fri, 2009-02-27 at 08:13 -0500, Larry Ludwig wrote:
> Hi all,
>
>
> Here's the latest test with devel, testing, production in memory
>
>
> stock 0.24.7
>
>
> PID USER PR NI VIRT RES SHR S %CPU %MEMTIME+ COMMAND
>
>
>
> 31650 puppet16 0 308m 254
Hi all,
Here's the latest test with devel, testing, production in memory
stock 0.24.7
PID USER PR NI VIRT RES SHR S %CPU %MEMTIME+ COMMAND
31650 puppet16 0 308m 254m 2864 S 0.3 8.2 29:09.94
puppetmasterd
31672 puppet16 0 304m 249m 2864 S 0.7 8.1 29:24.35
On Feb 26, 10:48 pm, Gary Law wrote:
> Diff from
> blastwave.rb follows.
> Gary
>
Note the diff is for illustrative purposes; not to be applied. This is
a new provider backend.
Gary
--~--~-~--~~~---~--~~
You received this message because you are subscribed to th
58 matches
Mail list logo
