On Mon, Sep 28, 2009 at 5:38 PM, Alexander Bodnarashik
<[email protected]> wrote:
> Thanks for the fix :)
> I've checked out
> http://hg.clusterlabs.org/pacemaker/stable-1.0/rev/05c8b63cbca7
> Now ports are open.
>
> I've encountered other problem though.
> I have 2 boxes in cluster - box1 and box2. Third box, not in cluster, is
> named farm.
> All of them are running Gentoo.

All have the same version of pacemaker?

> Cluster stack - openais-1.0.1
>
> Trying to issue cibadmin -Q from farm:
> 1234 - plain port
>>
>> CIB_server=box1.cluster CIB_port=1234 cibadmin -Q

So a couple of things here (that you couldn't possibly be expected to
know, sorry, i forgot to mention them at the time)...

You need to set CIB_user to the user than the remote node runs the CIB
as (eg. hacluster)
For plaintext connections, you need to set CIB_encrypted=false

Actually that first one needs to be the default (since non-root
daemons can only do PAM authentication for the user they're running
as).

>> Password:
>> cibadmin: Connection to box1.cluster:1234 failed:
>> Signon to CIB failed:
>> Init failed, could not perform requested operations
>
> and exits immediately
>
> 12345 - tls port
>>
>> CIB_server=box1.cluster CIB_port=12345 cibadmin -Q
>> Password:
>>
>
> and it freezes. In logs on box1 i can see following:
>>
>> Sep 28 18:03:30 box1 cib: [3342]: ERROR: crm_xml_err: XML Error: Entity:
>> line 1: parsererror : Start tag expected, '<' not found
>> Sep 28 18:03:30 box1 cib: [3342]: ERROR: crm_xml_err: XML Error:
>> Sep 28 18:03:30 box1 cib: [3342]: ERROR: crm_xml_err: XML Error: ^
>> Sep 28 18:03:30 box1 cib: [3342]: WARN: string2xml: Parsing failed
>> (domain=1, level=3, code=4): Start tag expected, '<' not found
>> Sep 28 18:03:30 box1 cib: [3342]: ERROR: string2xml: Couldn't parse 3
>> chars:
>> Sep 28 18:03:30 box1 cib: [3342]: ERROR: cib_recv_remote_msg: Couldn't
>> parse: ''
>
> After that i'm unable to run  on box1 neither crm_mon (writes Attempting
> connection to the cluster...) nor cibadmin -Q - it waits for a while and
> then writes following:
>>
>>  Signon to CIB failed: reply failed
>> Init failed, could not perform requested operations

Thats very disturbing.
Can you try running the CIB under valgrind to see if it reports
anything or interest?

export HA_VALGRIND_ENABLED=cib
export VALGRIND_OPTS="--log-file=/tmp/pacemaker-%p.valgrind
--leak-check=full --show-reachable=yes --trace-children=no
--num-callers=25"

install valgrind
then start the cluster

>
> On box2 crm_mon runs, but it doesn't reflect changes in cluster. running
> cibadmin -Q waits for a while, then shows following:
>>
>> Call cib_query failed (-41): Remote node did not respond
>> <null>
>
> Finally, in a few minutes i've found errors in logs (i think they are caused
> by my attempt to connect to cluster remotely), so attaching.
>
> Thanks.
>
> On Sep 21, 2009, at 13:53, Andrew Beekhof wrote:
>
>> I had a look at this, and basically I broke the initialization.
>> I'll fix this today for 1.0.6
>>
>> On Thu, Sep 10, 2009 at 10:05 PM, Andrew Beekhof <[email protected]>
>> wrote:
>>>
>>> Strange. I'll take a look on monday (after my vacation).
>>>
>>> On Wed, Sep 9, 2009 at 10:30 AM, Alexander
>>> Bodnarashik<[email protected]> wrote:
>>>>
>>>> On Sep 08, 2009, at 09:26, Andrew Beekhof wrote:
>>>>
>>>>> On Fri, Sep 4, 2009 at 5:35 PM, Alexander
>>>>> Bodnarashik<[email protected]>
>>>>> wrote:
>>>>>>
>>>>>> Hi. I'm trying to enable remote connections to cluster, but with no
>>>>>> luck, netstat does not show those ports as opened, logs tell me
>>>>>> nothing as well.
>>>>>
>>>>> Were those port values in the CIB when the cluster started?  If not,
>>>>> restart the cluster software.
>>>>> Otherwise, check if TLS support was enabled when you built pacemaker.
>>>>
>>>> Both port values were set before cluster started.
>>>>
>>>> I didn't find tls-related options in pacemaker "./configure". But tls
>>>> was
>>>> found  on system during configure script run:
>>>> ...
>>>> checking gnutls/gnutls.h usability... yes
>>>> checking gnutls/gnutls.h presence... yes
>>>> checking for gnutls/gnutls.h... yes
>>>> checking for security/pam_appl.h... (cached) yes
>>>> checking for pam/pam_appl.h... (cached) no
>>>> checking for libgnutls-config... /usr/bin/libgnutls-config
>>>> checking for gnutls header flags... -I/usr/include
>>>> checking for gnutls library flags... -L/usr/lib -lgnutls -lgcrypt
>>>> -lgpg-error
>>>> ...
>>>>
>>>> also cibadmin linked against gnutls:
>>>>  ldd `which cibadmin`|grep tls
>>>>       libgnutls.so.26 => /usr/lib/libgnutls.so.26 (0xb7fc5000)
>>>> So i suppose that tls is enabled.
>>>>
>>>> I'm also attaching logs, corosync config and cib.
>>>> Thanks.
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Linux-HA mailing list
>>>> [email protected]
>>>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>>>> See also: http://linux-ha.org/ReportingProblems
>>>>
>>>
>> _______________________________________________
>> Linux-HA mailing list
>> [email protected]
>> http://lists.linux-ha.org/mailman/listinfo/linux-ha
>> See also: http://linux-ha.org/ReportingProblems
>
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to