There isn't a perfect solution for this problem but only a workaround.
You must wrote a LSB init script that disables the VG's that heartbeat manage.
This script must start between lvm2 and heartbeat and use a simple command like:
/sbin/vgchange -an $VG
Obviously you must fill the variable $VG.

I hope this will help.

Bye.

Il giorno 29/lug/2010, alle ore 17.43, Alessandra Giovanardi ha scritto:

> Hi,
> 
> I have a cluster of two nodes (fec01 and fec02), with 2 resouce groups 
> (the first normally running on fec01 and the second one normally running 
> on fec02).
> The OS is Debian Lenny (stable).
> 
> My heartbeat version is:
> ii  heartbeat                                  2.1.3-6lenny4             
> ii  heartbeat-2                                2.1.3-6lenny4             
> ii  heartbeat-2-gui                            2.1.3-6lenny4             
> ii  heartbeat-gui                              2.1.3-6lenny4             
> 
> Hi have this problem:
> 
> If I stop heartbeat on node fec01, all services migrate on node fec02 
> (this is ok!).
> Furthermore if I reboot node fec01 (at boot I have heartbeat not 
> active), and then later I start heartbeat, on fec02 we have a quite 
> strange behaviour:
> 
> - services are stopped for a moment and then immediatly reactivated.
> 
> I have this log error:
> mgmtd[24082]: 2010/07/29_17:21:43 ERROR: native_add_running: Resource 
> ocf::LVM:resource_VG_METEO appears to be active on 2 nodes.
> mgmtd[24082]: 2010/07/29_17:21:43 ERROR: See 
> http://linux-ha.org/v2/faq/resource_too_active for more information.
> mgmtd[24082]: 2010/07/29_17:21:43 ERROR: native_add_running: Resource 
> ocf::LVM:resource_VG_WEBS appears to be active on 2 nodes.
> mgmtd[24082]: 2010/07/29_17:21:43 ERROR: See 
> http://linux-ha.org/v2/faq/resource_too_active for more information.
> 
> I noted that at boot, fec01 has the LVMs, normally managed by heartbeat, 
> in ACTIVE status:
> fec01:~# lvscan
>  ACTIVE            '/dev/fec_webs_vg/fec_webs_lv' [499.00 GB] inherit
>  ACTIVE            '/dev/fec_meteo_vg/fec_meteo_lv' [199.00 GB] inherit
> 
> In SUSE the problem can be solved by customizing this variable:
> # cat /etc/sysconfig/lvm | grep -v ^#
> LVM_VGS_ACTIVATED_ON_BOOT=""
> 
> In Debian I don't find any similar setting.
> 
> Could you help me to solve the problem?
> 
> In fact we are usually migrating resource from a node to another (to 
> upgrade OS) and then perform reboot (and we would not have this services 
> flap problem).
> 
> I send you also this old post:
> http://lists.community.tummy.com/pipermail/linux-ha/2009-April/036842.html
> 
> Thanks
> A.
> 
> -- 
> Alessandra Giovanardi
> e-mail [email protected]
> Cineca - DSET http://www.cineca.it
> Phone +39 051 6171939
> 
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems


Michele Codutti
Centro Servizi Informatici e Telematici (CSIT)
Universita' degli Studi di Udine
via Delle Scienze, 208 - 33100 UDINE
tel +39 0432 558928
fax +39 0432 558911
e-mail: michele.codutti at uniud.it

_______________________________________________
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