One way I have seen for setting up system users is to create them virtually and then realize them with the spaceship operator, say by ggroup -- like this:
In the referenced posting, the OP was trying to create a bunch of virtual users with all the parameter data in hiera.
But this is usually done with create_resources and that don't do virtual :(
So, I am asking a couple of questions going in different directions:
Can anyone show me a way to create a virtual resource from hiera ?
Actually, that should be: Hot to create a bunch of virtual resources from a hiera-hash
Contrarywise, it was suggested in the referenced posting that one
"could find all users in all hierarchies with hiera_hash and then declare them at once".
The down side of that is that using the spaceship operator,
one can filter by class parameters and there is no problem if the same virtual resource is realized more than once.
I am not certain that the hash-merge would produce a list of unique resource parameters.
If the same user was in more than one hiera file, would it appear multiple times in the hash ?
On Thu, May 17, 2012 at 9:12 PM, Dan White <y...@comcast.net> wrote:
> One way I have seen for setting up system users is to create them
> virtually and then realize them with the spaceship operator, say by ggroup
> -- like this:
> In the referenced posting, the OP was trying to create a bunch of virtual
> users with all the parameter data in hiera.
> But this is usually done with create_resources and that don't do virtual :(
> So, I am asking a couple of questions going in different directions:
> Can anyone show me a way to create a virtual resource from hiera ?
> Actually, that should be: Hot to create a bunch of virtual resources from
> a hiera-hash
> Contrarywise, it was suggested in the referenced posting that one
> "could find all users in all hierarchies with hiera_hash and then declare
> them at once".
> The down side of that is that using the spaceship operator,
> one can filter by class parameters and there is no problem if the same
> virtual resource is realized more than once.
> I am not certain that the hash-merge would produce a list of unique
> resource parameters.
> If the same user was in more than one hiera file, would it appear multiple
> times in the hash ?
That's exactly what I was suggesting. If you declared the users as
create_resources expected it, and included the same user in multiple levels
of the hierarchy, the hash merging functionality would leave you with a
single user declaration. One caveat (as of right now): you need to declare
ALL the parameters in ALL the levels of the hierarchy.
> --
> You received this message because you are subscribed to the Google Groups
> "Puppet Users" group.
> To post to this group, send email to puppet-users@googlegroups.com.
> To unsubscribe from this group, send email to
> puppet-users+unsubscribe@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/puppet-users?hl=en.
--
Gary Larizza
Professional Services Engineer
Puppet Labs
> On Thu, May 17, 2012 at 9:12 PM, Dan White <y...@comcast.net> wrote:
> > One way I have seen for setting up system users is to create them
> > virtually and then realize them with the spaceship operator, say by ggroup
> > -- like this:
> > In the referenced posting, the OP was trying to create a bunch of virtual
> > users with all the parameter data in hiera.
> > But this is usually done with create_resources and that don't do virtual :(
> > So, I am asking a couple of questions going in different directions:
> > Can anyone show me a way to create a virtual resource from hiera ?
> > Actually, that should be: Hot to create a bunch of virtual resources from
> > a hiera-hash
> > Contrarywise, it was suggested in the referenced posting that one
> > "could find all users in all hierarchies with hiera_hash and then declare
> > them at once".
> > The down side of that is that using the spaceship operator,
> > one can filter by class parameters and there is no problem if the same
> > virtual resource is realized more than once.
> > I am not certain that the hash-merge would produce a list of unique
> > resource parameters.
> > If the same user was in more than one hiera file, would it appear multiple
> > times in the hash ?
> That's exactly what I was suggesting. If you declared the users as
> create_resources expected it, and included the same user in multiple levels
> of the hierarchy, the hash merging functionality would leave you with a
> single user declaration. One caveat (as of right now): you need to declare
> ALL the parameters in ALL the levels of the hierarchy.
> --
Ah, Gary ! Just the brain I wish to pick ! :)
Thanks for the response. It makes sense.
However, if I perceive this properly, it would provide an All-Or-Nothing implementation of users.
I am looking for for a way to have a master list of users in hiera and then realize/instantiate a group-keyed-subset of the master list on each node.
“Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.”
Bill Waterson (Calvin & Hobbes)
> Thanks for the response. It makes sense.
> However, if I perceive this properly, it would provide an All-Or-Nothing implementation of users.
> I am looking for for a way to have a master list of users in hiera and then realize/instantiate a group-keyed-subset of the master list on each node.
> Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.
> Bill Waterson (Calvin& Hobbes)
Why not use a define to wrap the virtualised declaration of your users, or tried and doesn't work? Theoretically like this:
Information Systems
Ph: +44 (0) 20 3192 2520
luke.bi...@lmax.com | http://www.lmax.com LMAX, Yellow Building, 1A Nicholas Road, London W11 4AN
FX and CFDs are leveraged products that can result in losses exceeding
your deposit. They are not suitable for everyone so please ensure you
fully understand the risks involved. The information in this email is not
directed at residents of the United States of America or any other
jurisdiction where trading in CFDs and/or FX is restricted or prohibited
by local laws or regulations.
The information in this email and any attachment is confidential and is
intended only for the named recipient(s). The email may not be disclosed
or used by any person other than the addressee, nor may it be copied in
any way. If you are not the intended recipient please notify the sender
immediately and delete any copies of this message. Any unauthorised
copying, disclosure or distribution of the material in this e-mail is
strictly forbidden.
LMAX operates a multilateral trading facility. Authorised and regulated by the Financial Services Authority (firm registration number 509778) and
is registered in England and Wales (number 06505809). Our registered address is Yellow Building, 1A Nicholas Road, London, W11
4AN.
> On 18/05/12 13:46, Dan White wrote:
> > Ah, Gary ! Just the brain I wish to pick ! :)
> > Thanks for the response. It makes sense.
> > However, if I perceive this properly, it would provide an All-Or-Nothing implementation of users.
> > I am looking for for a way to have a master list of users in hiera and then realize/instantiate a group-keyed-subset of the master list on each node.
> > “Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.”
> > Bill Waterson (Calvin& Hobbes)
> Why not use a define to wrap the virtualised declaration of your users, > or tried and doesn't work? Theoretically like this:
... and puppet screamed "could not create resource of unknown type @users::mkuser"
I can certainly try that and see what happens.
Thanks.
“Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.”
Bill Waterson (Calvin & Hobbes)
On May 17, 11:16 pm, Gary Larizza <g...@puppetlabs.com> wrote:
> That's exactly what I was suggesting. If you declared the users as
> create_resources expected it, and included the same user in multiple levels
> of the hierarchy, the hash merging functionality would leave you with a
> single user declaration. One caveat (as of right now): you need to declare
> ALL the parameters in ALL the levels of the hierarchy.
That's because hiera hash merging is shallow, right? That is, every
value in the resulting hash is the exact one given for the associated
key at some level of the hierarchy? I'm not complaining; I just want
to confirm that I understand correctly.
> On May 17, 11:16 pm, Gary Larizza <g...@puppetlabs.com <javascript:;>>
> wrote:
> > That's exactly what I was suggesting. If you declared the users as
> > create_resources expected it, and included the same user in multiple
> levels
> > of the hierarchy, the hash merging functionality would leave you with a
> > single user declaration. One caveat (as of right now): you need to
> declare
> > ALL the parameters in ALL the levels of the hierarchy.
> That's because hiera hash merging is shallow, right? That is, every
> value in the resulting hash is the exact one given for the associated
> key at some level of the hierarchy? I'm not complaining; I just want
> to confirm that I understand correctly.
Yep, Hiera doesn't currently support deep merges.
> John
> --
> You received this message because you are subscribed to the Google Groups
> "Puppet Users" group.
> To post to this group, send email to puppet-users@googlegroups.com<javascript:;>
> .
> To unsubscribe from this group, send email to
> puppet-users+unsubscribe@googlegroups.com <javascript:;>.
> For more options, visit this group at
> http://groups.google.com/group/puppet-users?hl=en.
--
Gary Larizza
Professional Services Engineer
Puppet Labs
> On 18/05/12 13:46, Dan White wrote:
> > Ah, Gary ! Just the brain I wish to pick ! :)
> > Thanks for the response. It makes sense.
> > However, if I perceive this properly, it would provide an All-Or-Nothing implementation of users.
> > I am looking for for a way to have a master list of users in hiera and then realize/instantiate a group-keyed-subset of the master list on each node.
> > “Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.”
> > Bill Waterson (Calvin& Hobbes)
> Why not use a define to wrap the virtualised declaration of your users, > or tried and doesn't work? Theoretically like this:
-------------------------
/etc/puppet/hierdata/common.yaml:
-------------------------
users: beast:
v_username : beast
v_uid : 6666
v_ingroups : v_info : Let's see if this works
boo:
v_username : boo
v_uid : 6667
v_ingroups : v_info : Let's see if this works also
-------------------------
And if I say:
hiera users -c /etc/puppet/hiera.yaml
I get:
{"beast"=>{"v_info"=>"Let's see if this works", "v_username"=>"beast", "v_ingroups"=>nil, "v_uid"=>6666}, "boo"=>{"v_info"=>"Let's see if this works also", "v_username"=>"boo", "v_ingroups"=>nil, "v_uid"=>6667}}
Which is out of order, but everything is there. That's how hashes are supposed to work !
But when I try to run the catalog with $the_users = hiera('users') or $the_users = hiera_hash('users')
followed by :
create_resources ( 'add_virtual_user', $the_users )
I get:
Must pass a parameter or all necessary values at /etc/puppet/manifests/nodes/puppetmaster-node.pp:15 on node puppetmaster.foo.org
“Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.”
Bill Waterson (Calvin & Hobbes)
On Fri, May 18, 2012 at 2:41 PM, Dan White <y...@comcast.net> wrote:
> ----- Luke Bigum <Luke.Bi...@lmax.com> wrote:
> > On 18/05/12 13:46, Dan White wrote:
> > > Ah, Gary ! Just the brain I wish to pick ! :)
> > > Thanks for the response. It makes sense.
> > > However, if I perceive this properly, it would provide an
> All-Or-Nothing implementation of users.
> > > I am looking for for a way to have a master list of users in hiera and
> then realize/instantiate a group-keyed-subset of the master list on each
> node.
> > > “Sometimes I think the surest sign that intelligent life exists
> elsewhere in the universe is that none of it has tried to contact us.”
> > > Bill Waterson (Calvin& Hobbes)
> > Why not use a define to wrap the virtualised declaration of your users,
> > or tried and doesn't work? Theoretically like this:
> define add_virtual_user ( $v_username, $v_uid, $v_ingroups, $v_info ) {
> @add_user { "${name}":
> username => "${v_username}",
> uid => "${v_uid}",
> ingroups => "${v_ingroups}",
> info => "${v_info}",
> }
> }
> -------------------------
> /etc/puppet/hierdata/common.yaml:
> -------------------------
> users:
> beast:
> v_username : beast
> v_uid : 6666
> v_ingroups :
> v_info : Let's see if this works
> boo:
> v_username : boo
> v_uid : 6667
> v_ingroups :
> v_info : Let's see if this works also
> -------------------------
> And if I say:
> hiera users -c /etc/puppet/hiera.yaml
> I get:
> {"beast"=>{"v_info"=>"Let's see if this works", "v_username"=>"beast",
> "v_ingroups"=>nil, "v_uid"=>6666}, "boo"=>{"v_info"=>"Let's see if this
> works also", "v_username"=>"boo", "v_ingroups"=>nil, "v_uid"=>6667}}
> Which is out of order, but everything is there. That's how hashes are
> supposed to work !
> But when I try to run the catalog with
> $the_users = hiera('users') or $the_users = hiera_hash('users')
> followed by :
> create_resources ( 'add_virtual_user', $the_users )
> I get:
> Must pass a parameter or all necessary values at
> /etc/puppet/manifests/nodes/puppetmaster-node.pp:15 on node
> puppetmaster.foo.org
> “Sometimes I think the surest sign that intelligent life exists elsewhere
> in the universe is that none of it has tried to contact us.”
> Bill Waterson (Calvin & Hobbes)
> --
> You received this message because you are subscribed to the Google Groups
> "Puppet Users" group.
> To post to this group, send email to puppet-users@googlegroups.com.
> To unsubscribe from this group, send email to
> puppet-users+unsubscribe@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/puppet-users?hl=en.
--
Gary Larizza
Professional Services Engineer
Puppet Labs
I agree with Gary, Dan, it's probably the lack of data in the 'v_ingroups' key in your YAML that create_resources() is complaining about. If it truly can't pass an empty key/val pair you could do something hacky like use the string "undef" then explicitly check for it in the define.
> I wonder if the nil value is not being accepted as 'being passed' - > can you populate the nil value and try again?
> On Fri, May 18, 2012 at 2:41 PM, Dan White <y...@comcast.net > <mailto:y...@comcast.net>> wrote:
> ----- Luke Bigum <Luke.Bi...@lmax.com
> <mailto:Luke.Bi...@lmax.com>> wrote:
> > On 18/05/12 13:46, Dan White wrote:
> > > Ah, Gary ! Just the brain I wish to pick ! :)
> > > Thanks for the response. It makes sense.
> > > However, if I perceive this properly, it would provide an
> All-Or-Nothing implementation of users.
> > > I am looking for for a way to have a master list of users in
> hiera and then realize/instantiate a group-keyed-subset of the
> master list on each node.
> > > �Sometimes I think the surest sign that intelligent life
> exists elsewhere in the universe is that none of it has tried to
> contact us.�
> > > Bill Waterson (Calvin& Hobbes)
> > Why not use a define to wrap the virtualised declaration of your
> users,
> > or tried and doesn't work? Theoretically like this:
> define add_virtual_user ( $v_username, $v_uid, $v_ingroups,
> $v_info ) {
> @add_user { "${name}":
> username => "${v_username}",
> uid => "${v_uid}",
> ingroups => "${v_ingroups}",
> info => "${v_info}",
> }
> }
> -------------------------
> /etc/puppet/hierdata/common.yaml:
> -------------------------
> users:
> beast:
> v_username : beast
> v_uid : 6666
> v_ingroups :
> v_info : Let's see if this works
> boo:
> v_username : boo
> v_uid : 6667
> v_ingroups :
> v_info : Let's see if this works also
> -------------------------
> And if I say:
> hiera users -c /etc/puppet/hiera.yaml
> I get:
> {"beast"=>{"v_info"=>"Let's see if this works",
> "v_username"=>"beast", "v_ingroups"=>nil, "v_uid"=>6666},
> "boo"=>{"v_info"=>"Let's see if this works also",
> "v_username"=>"boo", "v_ingroups"=>nil, "v_uid"=>6667}}
> Which is out of order, but everything is there. That's how hashes
> are supposed to work !
> But when I try to run the catalog with
> $the_users = hiera('users') or $the_users = hiera_hash('users')
> followed by :
> create_resources ( 'add_virtual_user', $the_users )
> I get:
> Must pass a parameter or all necessary values at
> /etc/puppet/manifests/nodes/puppetmaster-node.pp:15 on node
> puppetmaster.foo.org <http://puppetmaster.foo.org>
> �Sometimes I think the surest sign that intelligent life exists
> elsewhere in the universe is that none of it has tried to contact us.�
> Bill Waterson (Calvin & Hobbes)
> --
> You received this message because you are subscribed to the Google
> Groups "Puppet Users" group.
> To post to this group, send email to puppet-users@googlegroups.com
> <mailto:puppet-users@googlegroups.com>.
> To unsubscribe from this group, send email to
> puppet-users+unsubscribe@googlegroups.com
> <mailto:puppet-users%2Bunsubscribe@googlegroups.com>.
> For more options, visit this group at
> http://groups.google.com/group/puppet-users?hl=en.
> --
> Gary Larizza
> Professional Services Engineer
> Puppet Labs
> -- > You received this message because you are subscribed to the Google > Groups "Puppet Users" group.
> To post to this group, send email to puppet-users@googlegroups.com.
> To unsubscribe from this group, send email to > puppet-users+unsubscribe@googlegroups.com.
> For more options, visit this group at > http://groups.google.com/group/puppet-users?hl=en.
-- Luke Bigum
Information Systems
Ph: +44 (0) 20 3192 2520
luke.bi...@lmax.com | http://www.lmax.com LMAX, Yellow Building, 1A Nicholas Road, London W11 4AN
FX and CFDs are leveraged products that can result in losses exceeding
your deposit. They are not suitable for everyone so please ensure you
fully understand the risks involved. The information in this email is not
directed at residents of the United States of America or any other
jurisdiction where trading in CFDs and/or FX is restricted or prohibited
by local laws or regulations.
The information in this email and any attachment is confidential and is
intended only for the named recipient(s). The email may not be disclosed
or used by any person other than the addressee, nor may it be copied in
any way. If you are not the intended recipient please notify the sender
immediately and delete any copies of this message. Any unauthorised
copying, disclosure or distribution of the material in this e-mail is
strictly forbidden.
LMAX operates a multilateral trading facility. Authorised and regulated by the Financial Services Authority (firm registration number 509778) and
is registered in England and Wales (number 06505809). Our registered address is Yellow Building, 1A Nicholas Road, London, W11
4AN.
On Mon, May 21, 2012 at 1:24 AM, Luke Bigum <Luke.Bi...@lmax.com> wrote:
> I agree with Gary, Dan, it's probably the lack of data in the
> 'v_ingroups' key in your YAML that create_resources() is complaining about.
> If it truly can't pass an empty key/val pair you could do something hacky
> like use the string "undef" then explicitly check for it in the define.
> On Mon, May 21, 2012 at 1:24 AM, Luke Bigum <Luke.Bi...@lmax.com > <mailto:Luke.Bi...@lmax.com>> wrote:
> I agree with Gary, Dan, it's probably the lack of data in the
> 'v_ingroups' key in your YAML that create_resources() is
> complaining about. If it truly can't pass an empty key/val pair
> you could do something hacky like use the string "undef" then
> explicitly check for it in the define.
> Do you really mean to be comparing to the string "undef" rather than > the keyword undef (no quotes)?
Yes, unfortunately I did. It's because when using Hiera 0.3 it's a bit difficult to figure out what a Ruby nil gets passed into Puppet as. Consider the following manifest using Dan's example YAML (v_ingroups is a nil value):
----------------
#---
#users:
# beast:
# v_username : beast
# v_uid : 6666
# v_ingroups :
# v_info : Let's see if this works
It's not an empty string, it's not undef (but when you print it it comes out as undef), it's not nil (which doesn't exist in Puppet), it's not false but it *is* true? I've came across this once before and can't remember what "nil" actually gets interpreted as.
So if you feed that Puppet hash directly into the create_resources() function, it complains about a missing parameter:
---------------------
biguml@biguml-laptop:~$ puppet apply test.pp
notice: Scope(Class[main]): v_usernamebeastv_uid6666v_ingroupsundefv_infoLet's see if this works
notice: Scope(Class[main]): undef
notice: Scope(Class[main]): is true
Must pass a parameter or all necessary values at /home/biguml/test.pp:40 on node biguml-laptop
---------------------
So my suggestion was to explicitly set "undef" as a string in the yaml, then match on that in the Puppet manifests. It's horrible but would work.
> If you want to test if a variable is undefined the best way is to do this:
> if ($foo == undef) { notice "\$foo is undef" }
> else { notice "\$foo is defined as ${foo}" }
> -Jeff
> -- > You received this message because you are subscribed to the Google > Groups "Puppet Users" group.
> To post to this group, send email to puppet-users@googlegroups.com.
> To unsubscribe from this group, send email to > puppet-users+unsubscribe@googlegroups.com.
> For more options, visit this group at > http://groups.google.com/group/puppet-users?hl=en.
-- Luke Bigum
Information Systems
Ph: +44 (0) 20 3192 2520
luke.bi...@lmax.com | http://www.lmax.com LMAX, Yellow Building, 1A Nicholas Road, London W11 4AN
FX and CFDs are leveraged products that can result in losses exceeding
your deposit. They are not suitable for everyone so please ensure you
fully understand the risks involved. The information in this email is not
directed at residents of the United States of America or any other
jurisdiction where trading in CFDs and/or FX is restricted or prohibited
by local laws or regulations.
The information in this email and any attachment is confidential and is
intended only for the named recipient(s). The email may not be disclosed
or used by any person other than the addressee, nor may it be copied in
any way. If you are not the intended recipient please notify the sender
immediately and delete any copies of this message. Any unauthorised
copying, disclosure or distribution of the material in this e-mail is
strictly forbidden.
LMAX operates a multilateral trading facility. Authorised and regulated by the Financial Services Authority (firm registration number 509778) and
is registered in England and Wales (number 06505809). Our registered address is Yellow Building, 1A Nicholas Road, London, W11
4AN.
I found an answer to this particular issue. Thanks for the reminder so I can share the answer:
I found the hiera/yaml way to indicate an empty array !
So, to use my earlier example:
users:
beast:
username : beast
uid : 6666
ingroups :
- ''
info : Let's see if this works
Then, with a hiera call, I get :
{"beast"=>{"ingroups"=>[""], "uid"=>6666, "username"=>"beast", "info"=>"Let's see if this works"}
I was able to more forward past this problem after figuring that out.
“Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.”
Bill Waterson (Calvin & Hobbes)
> On 22/05/12 00:22, Jeff McCune wrote:
> > On Mon, May 21, 2012 at 1:24 AM, Luke Bigum <Luke.Bi...@lmax.com > > <mailto:Luke.Bi...@lmax.com>> wrote:
> > I agree with Gary, Dan, it's probably the lack of data in the
> > 'v_ingroups' key in your YAML that create_resources() is
> > complaining about. If it truly can't pass an empty key/val pair
> > you could do something hacky like use the string "undef" then
> > explicitly check for it in the define.
> > Do you really mean to be comparing to the string "undef" rather than > > the keyword undef (no quotes)?
> Yes, unfortunately I did. It's because when using Hiera 0.3 it's a bit > difficult to figure out what a Ruby nil gets passed into Puppet as. > Consider the following manifest using Dan's example YAML (v_ingroups is > a nil value):
> ----------------
> #---
> #users:
> # beast:
> # v_username : beast
> # v_uid : 6666
> # v_ingroups :
> # v_info : Let's see if this works
> It's not an empty string, it's not undef (but when you print it it comes > out as undef), it's not nil (which doesn't exist in Puppet), it's not > false but it *is* true? I've came across this once before and can't > remember what "nil" actually gets interpreted as.
> So if you feed that Puppet hash directly into the create_resources() > function, it complains about a missing parameter:
> ---------------------
> biguml@biguml-laptop:~$ puppet apply test.pp
> notice: Scope(Class[main]): > v_usernamebeastv_uid6666v_ingroupsundefv_infoLet's see if this works
> notice: Scope(Class[main]): undef
> notice: Scope(Class[main]): is true
> Must pass a parameter or all necessary values at /home/biguml/test.pp:40 > on node biguml-laptop
> ---------------------
> So my suggestion was to explicitly set "undef" as a string in the yaml, > then match on that in the Puppet manifests. It's horrible but would work.
> -Luke
> > There's a big difference...
> > If you want to test if a variable is undefined the best way is to do this:
> > if ($foo == undef) { notice "\$foo is undef" }
> > else { notice "\$foo is defined as ${foo}" }
> > -Jeff
> > -- > > You received this message because you are subscribed to the Google > > Groups "Puppet Users" group.
> > To post to this group, send email to puppet-users@googlegroups.com.
> > To unsubscribe from this group, send email to > > puppet-users+unsubscribe@googlegroups.com.
> > For more options, visit this group at > > http://groups.google.com/group/puppet-users?hl=en.
> -- > Luke Bigum
> Information Systems
> Ph: +44 (0) 20 3192 2520
> luke.bi...@lmax.com | http://www.lmax.com > LMAX, Yellow Building, 1A Nicholas Road, London W11 4AN
> FX and CFDs are leveraged products that can result in losses exceeding
> your deposit. They are not suitable for everyone so please ensure you
> fully understand the risks involved. The information in this email is not
> directed at residents of the United States of America or any other
> jurisdiction where trading in CFDs and/or FX is restricted or prohibited
> by local laws or regulations.
> The information in this email and any attachment is confidential and is
> intended only for the named recipient(s). The email may not be disclosed
> or used by any person other than the addressee, nor may it be copied in
> any way. If you are not the intended recipient please notify the sender
> immediately and delete any copies of this message. Any unauthorised
> copying, disclosure or distribution of the material in this e-mail is
> strictly forbidden.
> LMAX operates a multilateral trading facility. Authorised and regulated > by the Financial Services Authority (firm registration number 509778) and
> is registered in England and Wales (number 06505809). > Our registered address is Yellow Building, 1A Nicholas Road, London, W11
> 4AN.
> -- > You received this message because you are subscribed to the Google Groups "Puppet Users" group.
> To post to this group, send email to puppet-users@googlegroups.com.
> To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
> I was able to more forward past this problem after figuring that out.
> “Sometimes I think the surest sign that intelligent life exists elsewhere
> in the universe is that none of it has tried to contact us.”
> Bill Waterson (Calvin & Hobbes)
> ----- Luke Bigum <Luke.Bi...@lmax.com <javascript:;>> wrote:
> > On 22/05/12 00:22, Jeff McCune wrote:
> > > On Mon, May 21, 2012 at 1:24 AM, Luke Bigum <Luke.Bi...@lmax.com<javascript:;>
> > > <mailto:Luke.Bi...@lmax.com>> wrote:
> > > I agree with Gary, Dan, it's probably the lack of data in the
> > > 'v_ingroups' key in your YAML that create_resources() is
> > > complaining about. If it truly can't pass an empty key/val pair
> > > you could do something hacky like use the string "undef" then
> > > explicitly check for it in the define.
> > > Do you really mean to be comparing to the string "undef" rather than
> > > the keyword undef (no quotes)?
> > Yes, unfortunately I did. It's because when using Hiera 0.3 it's a bit
> > difficult to figure out what a Ruby nil gets passed into Puppet as.
> > Consider the following manifest using Dan's example YAML (v_ingroups is
> > a nil value):
> > ----------------
> > #---
> > #users:
> > # beast:
> > # v_username : beast
> > # v_uid : 6666
> > # v_ingroups :
> > # v_info : Let's see if this works
> > It's not an empty string, it's not undef (but when you print it it comes
> > out as undef), it's not nil (which doesn't exist in Puppet), it's not
> > false but it *is* true? I've came across this once before and can't
> > remember what "nil" actually gets interpreted as.
> > So if you feed that Puppet hash directly into the create_resources()
> > function, it complains about a missing parameter:
> > ---------------------
> > biguml@biguml-laptop:~$ puppet apply test.pp
> > notice: Scope(Class[main]):
> > v_usernamebeastv_uid6666v_ingroupsundefv_infoLet's see if this works
> > notice: Scope(Class[main]): undef
> > notice: Scope(Class[main]): is true
> > Must pass a parameter or all necessary values at /home/biguml/test.pp:40
> > on node biguml-laptop
> > ---------------------
> > So my suggestion was to explicitly set "undef" as a string in the yaml,
> > then match on that in the Puppet manifests. It's horrible but would work.
> > -Luke
> > > There's a big difference...
> > > If you want to test if a variable is undefined the best way is to do
> this:
> > > if ($foo == undef) { notice "\$foo is undef" }
> > > else { notice "\$foo is defined as ${foo}" }
> > > -Jeff
> > > --
> > > You received this message because you are subscribed to the Google
> > > Groups "Puppet Users" group.
> > > To post to this group, send email to puppet-users@
On Tue, May 22, 2012 at 8:50 AM, Jeff McCune <j...@puppetlabs.com> wrote:
> On Tuesday, May 22, 2012, Dan White wrote:
>> I found an answer to this particular issue. Thanks for the reminder so I
>> can share the answer:
>> I found the hiera/yaml way to indicate an empty array !
>> So, to use my earlier example:
>> users:
>> beast:
>> username : beast
>> uid : 6666
>> ingroups :
>> - ''
>> info : Let's see if this works
>> Then, with a hiera call, I get :
>> {"beast"=>{"ingroups"=>[""], "uid"=>6666, "username"=>"beast",
>> "info"=>"Let's see if this works"}
>> This is actually a non-empty array hat had one element, the empt string.
OK, I had a look today. Much of the behavior of hashes and arrays whose
elements are not defined has been resolved in Puppet 3.0.0rc. If you could
try that out it would help us make sure your problem has actually been
solved in Puppet.
As to how to specify an empty array as the value of a hash key using Hiera
and Puppet, this is the way:
---
username: beast
uid: 6666
ingroups: []
info: Let's see if this works
Notice it's just an empty set of square braces, no empty string.
> This clearly seems like a bug in puppet and how it is handling Hash
> values. I'll take a look more as soon as I get into the office.
It is a bug, luckily we've fixed it in Puppet 3.0.x. Please give the
release candidates a try.
define noundefparams($Consumer="abc", $indexer="def"){ if $Consumer { notify {"$Consumer":} }
}
_________________ ]# puppet apply hieranil.pp Error: Received incomplete information - no value provided for parameter indexer at /tmp/experiment/hieranil.pp:15 on node puppet-master Wrapped exception: Received incomplete information - no value provided for parameter indexer Error: Received incomplete information - no value provided for parameter indexer at /tmp/experiment/hieranil.pp:15 on node puppet-master
On Thursday, May 24, 2012 3:15:58 AM UTC+5:30, Jeff McCune wrote:
> On Tue, May 22, 2012 at 8:50 AM, Jeff McCune <je...@puppetlabs.com<javascript:> > > wrote:
>> On Tuesday, May 22, 2012, Dan White wrote:
>>> I found an answer to this particular issue. Thanks for the reminder so >>> I can share the answer:
>>> I found the hiera/yaml way to indicate an empty array ! >>> So, to use my earlier example:
>>> users: >>> beast: >>> username : beast >>> uid : 6666 >>> ingroups : >>> - '' >>> info : Let's see if this works
>>> Then, with a hiera call, I get :
>>> {"beast"=>{"ingroups"=>[""], "uid"=>6666, "username"=>"beast", >>> "info"=>"Let's see if this works"}
>>> This is actually a non-empty array hat had one element, the empt string.
> OK, I had a look today. Much of the behavior of hashes and arrays whose > elements are not defined has been resolved in Puppet 3.0.0rc. If you could > try that out it would help us make sure your problem has actually been > solved in Puppet.
> As to how to specify an empty array as the value of a hash key using Hiera > and Puppet, this is the way:
> --- > username: beast > uid: 6666 > ingroups: [] > info: Let's see if this works
> Notice it's just an empty set of square braces, no empty string.
>> This clearly seems like a bug in puppet and how it is handling Hash >> values. I'll take a look more as soon as I get into the office.
> It is a bug, luckily we've fixed it in Puppet 3.0.x. Please give the > release candidates a try.
Great ! Now all I need is some deep hash merging...
“Sometimes I think the surest sign that intelligent life exists elsewhere in the universe is that none of it has tried to contact us.” Bill Waterson (Calvin & Hobbes)
----- Original Message -----
From: "Abhay" <chrungoo.ab...@gmail.com> To: puppet-users@googlegroups.com Sent: Thursday, February 14, 2013 9:15:38 AM Subject: Re: [Puppet Users] Hiera Questions: Virtual User Resources and Hiera
]# puppet apply hieranil.pp Error: Received incomplete information - no value provided for parameter indexer at /tmp/experiment/hieranil.pp:15 on node puppet-master Wrapped exception: Received incomplete information - no value provided for parameter indexer Error: Received incomplete information - no value provided for parameter indexer at /tmp/experiment/hieranil.pp:15 on node puppet-master
On Thursday, May 24, 2012 3:15:58 AM UTC+5:30, Jeff McCune wrote:
On Tue, May 22, 2012 at 8:50 AM, Jeff McCune < je...@puppetlabs.com > wrote:
<blockquote>
On Tuesday, May 22, 2012, Dan White wrote:
<blockquote>
I found an answer to this particular issue. Thanks for the reminder so I can share the answer:
I found the hiera/yaml way to indicate an empty array ! So, to use my earlier example:
users: beast: username : beast uid : 6666 ingroups : - '' info : Let's see if this works
Then, with a hiera call, I get :
{"beast"=>{"ingroups"=>[""], "uid"=>6666, "username"=>"beast", "info"=>"Let's see if this works"}
This is actually a non-empty array hat had one element, the empt string. </blockquote>
OK, I had a look today. Much of the behavior of hashes and arrays whose elements are not defined has been resolved in Puppet 3.0.0rc. If you could try that out it would help us make sure your problem has actually been solved in Puppet.
As to how to specify an empty array as the value of a hash key using Hiera and Puppet, this is the way:
--- username: beast uid: 6666 ingroups: [] info: Let's see if this works
Notice it's just an empty set of square braces, no empty string.
<blockquote>
This clearly seems like a bug in puppet and how it is handling Hash values. I'll take a look more as soon as I get into the office. </blockquote>
It is a bug, luckily we've fixed it in Puppet 3.0.x. Please give the release candidates a try.
-Jeff </blockquote>
-- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To unsubscribe from this group and stop receiving emails from it, send an email to puppet-users+unsubscribe@googlegroups.com. To post to this group, send email to puppet-users@googlegroups.com. Visit this group at http://groups.google.com/group/puppet-users?hl=en . For more options, visit https://groups.google.com/groups/opt_out .
