[Linux-HA] Learning the CIB
vini.bill at gmail.com
vini.bill at gmail.com
Tue Jan 9 07:51:06 MST 2007
Should I submit a bug?
Maybe a error is too much for a simple order problem. A warning would be
apreciated.
There are some attributes that I'm still looking for the possible values. As
you have recommended, I took a look at the /usr/lib/heartbeat/crm.dtd and
haven't found any of these:
- default_action_timeout
- stonith_action
what are the possible values for each one of them?
Vinicius Menezes
On 1/9/07, Andrew Beekhof <beekhof at gmail.com> wrote:
>
> On 1/8/07, vini.bill at gmail.com <vini.bill at gmail.com> wrote:
> > Ok, ok... Now I've fixed most part of the errors but I still theses and
> I'm
> > working on them for more them a hour now. Please take a look at them:
> >
> > crm_verify[5457]: 2007/01/08_17:05:56 ERROR: Element
> cluster_property_set
> > content does not follow the DTD, expecting (rule* , attributes), got
> (nvpair
> > nvpair nvpair nvpair nvpair nvpair nvpair )
> >
> > crm_verify[5457]: 2007/01/08_17:05:56 ERROR: Element primitive content
> does
> > not follow the DTD, expecting (operations? , meta_attributes* ,
> > instance_attributes*), got (instance_attributes operations )
> >
> > crm_verify[5457]: 2007/01/08_17:05:56 ERROR: Element primitive content
> does
> > not follow the DTD, expecting (operations? , meta_attributes* ,
> > instance_attributes*), got (instance_attributes operations )
> >
> > crm_verify[5457]: 2007/01/08_17:05:56 ERROR: Element primitive content
> does
> > not follow the DTD, expecting (operations? , meta_attributes* ,
> > instance_attributes*), got (instance_attributes operations )
> >
> > crm_verify[5457]: 2007/01/08_17:05:56 ERROR: Element primitive content
> does
> > not follow the DTD, expecting (operations? , meta_attributes* ,
> > instance_attributes*), got (instance_attributes operations )
>
> notice anything about the order that the elements are listed?
> XML considers the order important (which is annoying actually since
> the CRM doesnt care)
>
> > crm_verify[5457]: 2007/01/08_17:05:56 ERROR: validate_with_dtd:xml.c CIB
> > does not validate against /usr/lib/heartbeat/crm.dtd
> > crm_verify[5457]: 2007/01/08_17:05:56 ERROR: main:crm_verify.c CIB did
> not
> > pass DTD validation
> > crm_verify[5457]: 2007/01/08_17:05:56 ERROR: main:crm_verify.c Triggered
> > non-fatal assert at crm_verify.c:242 : crm_mem_stats(NULL) == FALSE
> > Errors found during check: config not valid
> > -V may provide more details
> >
> > So... What's wrong here? According to the docs ( the cib-annotated )
> there
>
> as mentioned elsewhere, that page is somewhat out of date.
> look for: /usr/lib/heartbeat/crm.dtd for the DTD appropriate to your
> version.
>
> > isn`t such a thing ( or I haven`t read it correctly ) as:
> >
> > <crm_config>
> > <cluster_property_set>
> > <rule>
> > <attribute>
> > ( the nvpairs come here )
> > </attribute>
> > </rule>
> > </cluster_property_set>
> > </crm_config>
> >
> > I'm going to complete 150 lines of cib.xml and I still could't make it
> run
> > only once.
> >
> > Vinicius Menezes
> >
> > On 1/8/07, Andrew Beekhof <beekhof at gmail.com> wrote:
> > >
> > > On 1/8/07, vini.bill at gmail.com <vini.bill at gmail.com> wrote:
> > > > indeed... a slash was missed on the nvpair directive =/. You read my
> > > mind! I
> > > > just copy and pasted it. Now I got some strange ( but not bizarre )
> > > errors.
> > > > Take a look at it:
> > > >
> > > >
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: do_id_check:xml.cDetected
> > > > <atributes.../> object without an ID. Assigned:
> > > > 860918f8-41f7-45c1-8647-6a9e8967f5a7
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: do_id_check:xml.cDetected
> > > > <instance_atributes.../> object without an ID. Assigned:
> > > > 7519f042-6c92-4043-973d-d3c025daf28f
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: do_id_check:xml.cDetected
> > > > <primitive.../> object without an ID. Assigned:
> > > > 0ecf7e01-7918-4e86-be84-23b77f64bef6
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: do_id_check:xml.cDetected
> > > > <instance_attributes.../> object without an ID. Assigned:
> > > > e699110b-3f3a-41ac-bfae-3b7b34883698
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: do_id_check:xml.cDetected
> > > > <atributes.../> object without an ID. Assigned:
> > > > 611a407b-9d1b-4bd0-bd50-dd0b42b7120b
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: do_id_check:xml.cDetected
> > > > <instance_atributes.../> object without an ID. Assigned:
> > > > 0a37ac6c-e662-4736-81a0-2a20aa6c207d
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: do_id_check:xml.cDetected
> > > > <primitive.../> object without an ID. Assigned:
> > > > 2a1f1310-e715-40f6-8d96-f7b1c6245255
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: do_id_check:xml.cDetected
> > > > <instance_attributes.../> object without an ID. Assigned:
> > > > 62ed87c7-08ae-43a0-8997-5916e868e5ba
> > > > crm_verify[5133]: 2007/01/08_15:05:13 ERROR: main:crm_verify.c ID
> Check
> > > > failed
> > > >
> > > > The CRMd is actually atributing default values... so... they're not
> > > actually
> > > > errors, but warnings right?
> > >
> > > no, they're pretty fatal (as in we wont load such a configuration at
> > > runtime)
> > >
> > > the reason is more apparent when you consider multiple machines all
> > > starting up and assigning their own values for "id"... then trying to
> > > figure out what is what
> > >
> > > >
> > > >
> > > > On 1/8/07, Andreas Kurz <andreas.kurz at gmail.com> wrote:
> > > > >
> > > > > Hello,
> > > > >
> > > > > You have some syntax errors in your cib:
> > > > >
> > > > > --- cib_vini.xml 2007-01-08 17:50:33.831573895 +0100
> > > > > +++ cib_vini.xml.org 2007-01-08 17:49:47.770369431 +0100
> > > > > @@ -30,7 +30,7 @@
> > > > > -->
> > > > > <instance_atributes>
> > > > > <atributes>
> > > > > - <nvpair id="IPaddr_1_ip"
> > > > > value="192.168.0.143" name="DB01_IPaddr"/>
> > > > > + <nvpair id="IPaddr_1_ip"
> > > > > value="192.168.0.143" name="DB01_IPaddr">
> > > > > </atributes>
> > > > > </instance_atributes>
> > > > > <!--
> > > > > @@ -84,7 +84,7 @@
> > > > > <primitive class="ocf" provider="hertbeat">
> > > > > <instance_atributes>
> > > > > <atributes>
> > > > > - <nvpair id="IPaddr_2_ip"
> > > > > value="192.168.0.143" name="DB02_IPaddr"/>
> > > > > + <nvpair id="IPaddr_2_ip"
> > > > > value="192.168.0.143" name="DB02_IPaddr">
> > > > > </atributes>
> > > > > </instance_atributes>
> > > > > <operations>
> > > > >
> > > > > Try running "xmllint cib_vini.xml" ... I like that tool for
> finding
> > > > > lost slashes and other cut & paste errors ;-)
> > > > >
> > > > > Regards,
> > > > > Andreas
> > > > >
> > > > >
> > > > > On 1/8/07, vini.bill at gmail.com <vini.bill at gmail.com> wrote:
> > > > > > sorry, really sorry. forgot to send the weird_output file.
> > > > > >
> > > > > > Vinicius Menezes
> > > > > >
> > > > > > On 1/8/07, vini.bill at gmail.com <vini.bill at gmail.com> wrote:
> > > > > > >
> > > > > > > Well... I've got one more question. I created a example file
> and
> > > seems
> > > > > to
> > > > > > > be correct ( it's the cib_vini.xml ). But when I try to
> verify it
> > > > > with
> > > > > > > "crm_verify --xml-file" I get really weird erros ( it's in the
> > > > > > > weird_output.txt ). I think it's weird because it says ( at
> the
> > > very
> > > > > first
> > > > > > > line ):
> > > > > > >
> > > > > > > "crm_verify[5019]: 2007/01/08_14:27:01 ERROR: parse_xml:
> > > xml.cExpected:
> > > > > > > nvpair"
> > > > > > >
> > > > > > > But it's impossible to define a operation with nvpair. What's
> > > wrong
> > > > > with
> > > > > > > it? Am I misunderstanding something?
> > > > > > >
> > > > > > > Vinicius Menezes
> > > > > > >
> > > > > > >
> > > > > > > But when I use the crm_verify I get really weird erros ( such
> as
> > > > > > >
> > > > > > > On 1/8/07, Andrew Beekhof < beekhof at gmail.com> wrote:
> > > > > > > >
> > > > > > > > On 1/8/07, vini.bill at gmail.com <vini.bill at gmail.com> wrote:
> > > > > > > > > Hi guys.
> > > > > > > > >
> > > > > > > > > Sorry for not sending anything ( my home HD died and I had
> to
> > > > > replace
> > > > > > > > it ),
> > > > > > > > > this weekend. Fortunantely what I had made untill now was
> a
> > > basic
> > > > > > > > > reorganization of this page:
> > > > > > > > >
> > > http://www.linux-ha.org/ClusterResourceManager/DTD1.0/Annotated ,
> > > > > and
> > > > > > > > not
> > > > > > > > > soo important.
> > > > > > > > >
> > > > > > > > > Now I intend to document parts of the CIB configuration
> > > options
> > > > > that
> > > > > > > > are
> > > > > > > > > somehow obscure throughout the Linux-HA documentation.
> So...
> > > my
> > > > > > > > questions
> > > > > > > > > is:
> > > > > > > > >
> > > > > > > > > - In the primitve directive we have a "type" property
> which
> > > should
> > > > > > > > receive
> > > > > > > > > some expected values. Wich are these values? I'm looking
> > > > > everywhere on
> > > > > > > > the
> > > > > > > > > Linux-HA site and still haven't found anything.
> > > > > > > >
> > > > > > > > its the "type" of resource you want. ie. IPaddr, FileSystem,
> > > etc.
> > > > > > > > there is no complete list because people can write their own
> > > > > > > >
> > > > > > > > when class=ocf: anything in
> /usr/lib/ocf/resource.d//$provider/
> > > > > > > > when class=lsb: anything in /etc/init.d
> > > > > > > > when class=heartbeat anything in /etc/ha.d/resource.d
> > > > > > > >
> > > > > > > > >
> > > > > > > > > Today I`ll probably ask more about the constraints, rules
> ans
> > > > > > > > expressions
> > > > > > > > > options.
> > > > > > > > >
> > > > > > > > > Vinicius Menezes
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > On 1/6/07, Andrew Beekhof < beekhof at gmail.com> wrote:
> > > > > > > > > >
> > > > > > > > > > On 1/5/07, Pavol Gono <palo.gono at gmail.com> wrote:
> > > > > > > > > > > Hi
> > > > > > > > > > >
> > > > > > > > > > > Nice reading is in source tree:
> > > > > > > > > > > crm/crm-1.0.dtd
> > > > > > > > > > > crm/crm.txt (chapter 5)
> > > > > > > > > > > Just take the latest sources from
> > > http://hg.linux-ha.org/dev/.
> > > > > Web
> > > > > > > > > > > page
> > > > > > > >
> http://www.linux-ha.org/ClusterResourceManager/DTD1.0/Annotated
> > > > > > > > > > > is maybe not quite up-to-date.
> > > > > > > > > > >
> > > > > > > > > > > I'll answer what is simple to answer :)
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > On 1/5/07, vini.bill at gmail.com <vini.bill at gmail.com>
> > > wrote:
> > > > > > > > > > > > Hi guys.
> > > > > > > > > > > >
> > > > > > > > > > > > As I said in a previous mail, I'm Writing a
> document
> > > > > > > > teaching how
> > > > > > > > > > to use
> > > > > > > > > > > > the CIB Tags, how to place them, what they mean and
> > > > > something
> > > > > > > > more
> > > > > > > > > > that I
> > > > > > > > > > > > haven't foreseen. I intend to contribute it back to
> the
> > > > > project.
> > > > > > > > I'd
> > > > > > > > > > like to
> > > > > > > > > > > > discuss some topics that I haven't understood very
> well
> > > so
> > > > > as I
> > > > > > > > can
> > > > > > > > > > write
> > > > > > > > > > > > about them. I'll also make some minor comparsions
> with
> > > XHTML
> > > > > to
> > > > > > > > make
> > > > > > > > > > things
> > > > > > > > > > > > a bit clearer to me but I won't make these on the
> final
> > > > > text.
> > > > > > > > > > > >
> > > > > > > > > > > > Today, before leaving work I'll seek somewhere
> to
> > > post
> > > > > it so
> > > > > > > > you
> > > > > > > > > > can
> > > > > > > > > > > > take a look at my progress. It`ll be in a Wiki
> format
> > > since
> > > > > I`m
> > > > > > > > going
> > > > > > > > > > to
> > > > > > > > > > > > publish it on the wiki we have here and to make
> easier
> > > to
> > > > > add to
> > > > > > > > the
> > > > > > > > > > main
> > > > > > > > > > > > documentation.
> > > > > > > > > > > >
> > > > > > > > > > > > 1) Is there any directive that isn't contained on
> the
> > > <cib>
> > > > > > > > </cib> ?
> > > > > > > > > > Can I
> > > > > > > > > > > > create groups of configurations ( if yes, what would
> be
> > > the
> > > > > > > > necessity
> > > > > > > > > > to
> > > > > > > > > > > > have it? ) ?
> > > > > > > > > >
> > > > > > > > > > i dont understand the question
> > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > 2) There are only 4 configuration tags: (1)
> constraints,
> > > (2)
> > > > > > > > > > crm_config and
> > > > > > > > > > > > (3) resources and the (4) nodes. Is that right ?
> > > > > > > > > > >
> > > > > > > > > > > yes
> > > > > > > > > >
> > > > > > > > > > correct. and how do we know this?
> > > > > > > > > > we know this because the DTD says so:
> > > > > > > > > > <!ELEMENT configuration (crm_config, nodes, resources,
> > > > > > > > constraints)>
> > > > > > > > > >
> > > > > > > > > > I suggest starting with wikipedia's entry on DTDs:
> > > > > > > > > >
> http://en.wikipedia.org/wiki/Document_Type_Definition
> > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > 3) What does the empty <nodes/> tag mean? The CRM
> daemon
> > > > > will
> > > > > > > > fill in
> > > > > > > > > > every
> > > > > > > > > > > > node automagically ?
> > > > > > > > > > >
> > > > > > > > > > > yes, user doesn't need to fill in nodes tag.
> > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > 4) What do operations do in a primitive?
> > > > > > > > > > >
> > > > > > > > > > > e.g. you can tell heartbeat to monitor your resource
> > > > > > > > > >
> > > > > > > > > > or override various defaults... like what action to take
> if
> > > a
> > > > > start
> > > > > > > > > > action fails (or how long to wait before deciding that
> it
> > > > > failed)
> > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > >
> > > > > > > > > > > > 5) What is the difference between <atributes> and
> > > > > > > > <instance_atributes>
> > > > > > > > > > ??
> > > > > > > > > >
> > > > > > > > > > more information?
> > > > > > > > > >
> > > > > > > > > > > > 6) Is there any tag missing on this page (
> > > > > > > > > > > >
> > > > > http://www.linux-ha.org/ClusterResourceManager/DTD1.0/Annotated)
> that
> > > > > > > > > > I
> > > > > > > > > > > > should look for by myself ?
> > > > > > > > > > >
> > > > > > > > > > > See crm/crm-1.0.dtd
> > > > > > > > > >
> > > > > > > > > > correct, the online version seems to be somewhat out of
> date
> > > > > > > > > >
> > > > > > > > > > > > 7) What does the empty <status/> tag mean?
> > > > > > > > > > >
> > > > > > > > > > > status tag is filled in by heartbeat at runtime, user
> > > > > shouldn't
> > > > > > > > write
> > > > > > > > > > > to it. The command "cibadmin -Ql" is good for reading
> this
> > > > > tag.
> > > > > > > > > > _______________________________________________
> > > > > > > > > > Linux-HA mailing list
> > > > > > > > > > Linux-HA at lists.linux-ha.org
> > > > > > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > > > > > > > > > See also: http://linux-ha.org/ReportingProblems
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > --
> > > > > > > > > ... Vinicius Menezes ...
> > > > > > > > > _______________________________________________
> > > > > > > > > Linux-HA mailing list
> > > > > > > > > Linux-HA at lists.linux-ha.org
> > > > > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > > > > > > > > See also: http://linux-ha.org/ReportingProblems
> > > > > > > > >
> > > > > > > > _______________________________________________
> > > > > > > > Linux-HA mailing list
> > > > > > > > Linux-HA at lists.linux-ha.org
> > > > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > > > > > > > See also: http://linux-ha.org/ReportingProblems
> > > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > ... Vinicius Menezes ...
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > ... Vinicius Menezes ...
> > > > > >
> > > > > >
> > > > > > _______________________________________________
> > > > > > Linux-HA mailing list
> > > > > > Linux-HA at lists.linux-ha.org
> > > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > > > > > See also: http://linux-ha.org/ReportingProblems
> > > > > >
> > > > > >
> > > > > >
> > > > > _______________________________________________
> > > > > Linux-HA mailing list
> > > > > Linux-HA at lists.linux-ha.org
> > > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > > > > See also: http://linux-ha.org/ReportingProblems
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > ... Vinicius Menezes ...
> > > > _______________________________________________
> > > > Linux-HA mailing list
> > > > Linux-HA at lists.linux-ha.org
> > > > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > > > See also: http://linux-ha.org/ReportingProblems
> > > >
> > > _______________________________________________
> > > Linux-HA mailing list
> > > Linux-HA at lists.linux-ha.org
> > > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > > See also: http://linux-ha.org/ReportingProblems
> > >
> >
> >
> >
> > --
> > ... Vinicius Menezes ...
> > _______________________________________________
> > Linux-HA mailing list
> > Linux-HA at lists.linux-ha.org
> > http://lists.linux-ha.org/mailman/listinfo/linux-ha
> > See also: http://linux-ha.org/ReportingProblems
> >
> _______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
--
... Vinicius Menezes ...
More information about the Linux-HA
mailing list