[Pacemaker] Problem : By colocations limitation, the resource appointment of the combination does not become effective.

Andrew Beekhof andrew at beekhof.net
Tue Mar 23 19:11:18 UTC 2010


2010/3/19  <renayama19661014 at ybb.ne.jp>:
> Hi Andrew,
>
>> I've been extremely busy.
>> Sometimes I defer more complex questions until I have time to give
>> them my full attention.
>
> I understand that you are busy.
> Thank you for comment.
>
>> I don't really understand the question here.
>
> Sorry..
> I made a mistake in the link of the former problem.
> I explain a problem sequentially once again.
>
> We constituted the next cluster.
>
> Online: [ srv01 srv02 srv03 srv04 ]
>
>  Resource Group: UMgroup01
>     UmVIPcheck (ocf::heartbeat:Dummy): Started srv01
>     UmIPaddr   (ocf::heartbeat:Dummy): Started srv01
>     UmDummy01  (ocf::heartbeat:Dummy): Started srv01
>     UmDummy02  (ocf::heartbeat:Dummy): Started srv01
>  Resource Group: OVDBgroup02-1
>     prmExPostgreSQLDB1 (ocf::heartbeat:Dummy): Started srv01
>     prmFsPostgreSQLDB1-1       (ocf::heartbeat:Dummy): Started srv01
>     prmFsPostgreSQLDB1-2       (ocf::heartbeat:Dummy): Started srv01
>     prmFsPostgreSQLDB1-3       (ocf::heartbeat:Dummy): Started srv01
>     prmIpPostgreSQLDB1 (ocf::heartbeat:Dummy): Started srv01
>     prmApPostgreSQLDB1 (ocf::heartbeat:Dummy): Started srv01
>  Resource Group: OVDBgroup02-2
>     prmExPostgreSQLDB2 (ocf::heartbeat:Dummy): Started srv02
>     prmFsPostgreSQLDB2-1       (ocf::heartbeat:Dummy): Started srv02
>     prmFsPostgreSQLDB2-2       (ocf::heartbeat:Dummy): Started srv02
>     prmFsPostgreSQLDB2-3       (ocf::heartbeat:Dummy): Started srv02
>     prmIpPostgreSQLDB2 (ocf::heartbeat:Dummy): Started srv02
>     prmApPostgreSQLDB2 (ocf::heartbeat:Dummy): Started srv02
>  Resource Group: OVDBgroup02-3
>     prmExPostgreSQLDB3 (ocf::heartbeat:Dummy): Started srv03
>     prmFsPostgreSQLDB3-1       (ocf::heartbeat:Dummy): Started srv03
>     prmFsPostgreSQLDB3-2       (ocf::heartbeat:Dummy): Started srv03
>     prmFsPostgreSQLDB3-3       (ocf::heartbeat:Dummy): Started srv03
>     prmIpPostgreSQLDB3 (ocf::heartbeat:Dummy): Started srv03
>     prmApPostgreSQLDB3 (ocf::heartbeat:Dummy): Started srv03
>  Resource Group: grpStonith1
>     prmStonithN1       (stonith:external/ssh): Started srv04
>  Resource Group: grpStonith2
>     prmStonithN2       (stonith:external/ssh): Started srv01
>  Resource Group: grpStonith3
>     prmStonithN3       (stonith:external/ssh): Started srv02
>  Resource Group: grpStonith4
>     prmStonithN4       (stonith:external/ssh): Started srv03
>  Clone Set: clnUMgroup01
>     Started: [ srv01 srv04 ]
>  Clone Set: clnPingd
>     Started: [ srv01 srv02 srv03 srv04 ]
>  Clone Set: clnDiskd1
>     Started: [ srv01 srv02 srv03 srv04 ]
>  Clone Set: clnG3dummy1
>     Started: [ srv01 srv02 srv03 srv04 ]
>  Clone Set: clnG3dummy2
>     Started: [ srv01 srv02 srv03 srv04 ]
>
> We encountered the problem that early resource placement did not obey location by this constitution.
>  * I asked next question before... http://www.gossamer-threads.com/lists/linuxha/pacemaker/60342
>
> This was a mistake of our setting.
>  (snip)
>      <rsc_colocation id="rsc_colocation01-1" rsc="UMgroup01" with-rsc="clnPingd" score="INFINITY"/>
>      <rsc_colocation id="rsc_colocation01-2" rsc="UMgroup01" with-rsc="clnPingd2" score="INFINITY"/>
>      <rsc_colocation id="rsc_colocation01-3" rsc="UMgroup01" with-rsc="clnUMgroup01"
> score="INFINITY"/>
>      <rsc_colocation id="rsc_colocation02-1-1" rsc="group02-1" with-rsc="clnPingd" score="INFINITY"/>
>      <rsc_colocation id="rsc_colocation02-1-2" rsc="group02-1" with-rsc="clnPingd2"
> score="INFINITY"/>
>      <rsc_colocation id="rsc_colocation02-2-1" rsc="group02-2" with-rsc="clnPingd" score="INFINITY"/>
>      <rsc_colocation id="rsc_colocation02-2-2" rsc="group02-2" with-rsc="clnPingd2"
> score="INFINITY"/>
>  (snip)
>
> And we set 1000 in colocation.
>
>  (snip)
>      <rsc_colocation id="rsc_colocation01-1" rsc="UMgroup01" with-rsc="clnPingd" score="1000"/>
>      <rsc_colocation id="rsc_colocation01-2" rsc="UMgroup01" with-rsc="clnPingd2" score="1000"/>
>      <rsc_colocation id="rsc_colocation01-3" rsc="UMgroup01" with-rsc="clnUMgroup01" score="1000"/>
>      <rsc_colocation id="rsc_colocation02-1-1" rsc="group02-1" with-rsc="clnPingd" score="1000"/>
>      <rsc_colocation id="rsc_colocation02-1-2" rsc="group02-1" with-rsc="clnPingd2" score="1000"/>
>      <rsc_colocation id="rsc_colocation02-2-1" rsc="group02-2" with-rsc="clnPingd" score="1000"/>
>      <rsc_colocation id="rsc_colocation02-2-2" rsc="group02-2" with-rsc="clnPingd2" score="1000"/>
>  (snip)
>
> Because we set 1000 in colocation, the resource was arranged in a node definitely.

Ok, but that wasn't what I was suggesting.

I was suggesting:

 <rsc_colocation id="rsc_colocation01-3" rsc="UMgroup01"
with-rsc="clnUMgroup01" score="INFINITY"/>

<rsc_location id="no-connectivity-01-1" rsc="UMgroup01">
   <rule id="clnPingd-exclude-rule" score="-INFINITY" boolean-op="or">
      <expression id="UMgroup01-clnPingd-exclude" attribute="clnPingd"
operation="not_defined"/>
      <expression id="UMgroup01-clnPingd-only-positive"
attribute="clnPingd" operation="lt" type="integer" value="1"/>
      <expression id="UMgroup01-clnPingd2-exclude"
attribute="clnPingd2" operation="not_defined"/>
      <expression id="UMgroup01-clnPingd2-only-positive"
attribute="clnPingd2" operation="lt" type="integer" value="1"/>
   </rule>
</rsc_location>

<rsc_location id="no-connectivity-02-1" rsc="group02-1">
   <rule idref="clnPingd-exclude-rule"/>
</rsc_location>

<rsc_location id="no-connectivity-02-1" rsc="group02-2">
   <rule idref="clnPingd-exclude-rule"/>
</rsc_location>



>
> We confirmed movement after the trouble of clnPingd by cluster constitution of this setting more.
> (The detailed procedure is an email of the beginnings of this matter.)
>
> But clnPingd does not start in srv01, but UMgroup01 starts after this.
> * Because there was colocation limitation, we did not expect start of UMgroup01.
>
> Your answer to solve this problem was to set INFINITY in colocation.
>
>> Only if you change:
>> <rsc_colocation id="rsc_colocation01-1" rsc="UMgroup01"
>> with-rsc="clnPingd" score="1000"/>
>>
>> to
>> <rsc_colocation id="rsc_colocation01-1" rsc="UMgroup01"
>> with-rsc="clnPingd" score="INFINITY"/>
>
> However, the early resource placement that we solved becomes invalid when I set colocation in
> INFINITY.
>
> By our cluster constitution, can you satisfy two next demands?
>
> 1)The resource placement of a right early cluster.
> 2)The start control of the resource of the combination by the colocation limitation.
>
> Is there setting of cib.xml to realize a demand?
>
> #I am not good at English.
> #Give me comment if you do not understand contents of my comment.
>
> Best Regards,
> Hideo Yamauchi.
>
>
> --- Andrew Beekhof <andrew at beekhof.net> wrote:
>
>> 2010/3/17  <renayama19661014 at ybb.ne.jp>:
>> > Hi Andrew,
>> >
>> > Please give my question an answer.
>>
>> I've been extremely busy.
>> Sometimes I defer more complex questions until I have time to give
>> them my full attention.
>>
>> >
>> > Best Regards,
>> > Hideo Yamauchi.
>> >
>> > --- renayama19661014 at ybb.ne.jp wrote:
>> >
>> >> Hi Andrew,
>> >>
>> >> Thank you for comment.
>> >>
>> >> I asked next question before.
>> >>
>> >>   http://www.gossamer-threads.com/lists/linuxha/pacemaker/61484
>> >>
>> >> I guessed from your this answer.
>> >> When I use cib.xml of the answer of before, is the limitation that it combined a start of
>> >> clnPingd
>> >> with after a node rebooted unrealizable?
>>
>> I don't really understand the question here.
>>
>> >>
>> >> Limitation of before :
>> >> >  <rsc_colocation id="rsc_colocation01-1" rsc="UMgroup01"
>> >> > with-rsc="clnPingd" score="1000"/>
>> >>
>> >> This limitation :
>> >> >  <rsc_colocation id="rsc_colocation01-1" rsc="UMgroup01"
>> >> > with-rsc="clnPingd" score="INFINITY"/>
>> >>
>> >> Is there a description method of cib.xml letting both limitation function at the same time?
>>
>> What would that achieve?
>>
>> _______________________________________________
>> Pacemaker mailing list
>> Pacemaker at oss.clusterlabs.org
>> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>
>
>
>
> _______________________________________________
> Pacemaker mailing list
> Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
>




More information about the Pacemaker mailing list