[Linux-HA] Re: Pengine does want to live any hints?

Dejan Muhamedagic dejanmm at fastmail.fm
Tue Sep 18 04:22:31 MDT 2007


Hi,

On Tue, Sep 18, 2007 at 10:02:43AM +0200, Robert Lindgren wrote:
> The problem was:
> #0 0xb7eaa4fd in group_color () from /usr/lib/libpengine.so.3
> #1 0xb7e9ce57 in stage5 () from /usr/lib/libpengine.so.3
> #2 0xb7e9c284 in do_calculations () from /usr/lib/libpengine.so.3
> #3 0xb7e9c733 in process_pe_message () from /usr/lib/libpengine.so.3
> #4 0xb7ed87c1 in subsystem_msg_dispatch () from /usr/lib/libcrmcommon.so.1
> #5 0xb7f153b6 in G_CH_dispatch_int () from /usr/lib/libplumb.so.1
> #6 0xb7e30df2 in g_main_context_dispatch () from /usr/lib/libglib-2.0.so.0
> This problem was caused by me removing the resources from a group. There was
> also a constraint on the group left in the cib. This caused pengine to die a
> horrid death.
> I consider it's a bug that one is allows to do this if the pengine is not
> able to survive the changes I make with cibadmin. Or it's a bug that
> cibadmin allows me to do nonsupported changes.

Whereas it arguably doesn't make sense to have an empty group,
pengine still should not segfault. Please file a bug.

Thanks,

Dejan

> BR
> Robert Lindgren
> 
> On 9/18/07, Robert Lindgren <robert.lindgren at gmail.com > wrote:
> >
> > Hi All,
> >
> > I have a two node cluster with two groups, one running mysql and one
> > running samba. I stopped the samba group and tried to remove it with
> > cibadmin
> >
> > cibadmin -D -o resources -X '<primitive id="R_fs_samba" class="ocf"
> > type="Filesystem" provider="heartbeat">'
> > cibadmin -D -o resources -X '<primitive id="R_samba" class="heartbeat"
> > type="samba" provider="heartbeat">'
> > cibadmin -D -o resources -X '<primitive id="R_drbd_samba"
> > class="heartbeat" type="drbddisk" provider="heartbeat">'
> > cibadmin -D -o resources -X '<primitive class="ocf" type="IPaddr2"
> > provider="heartbeat" id="R_192.168.12.196">'
> >
> > and now this happens all the time in the log:
> >
> > pengine[24820]: 2007/09/18_09:06:41 WARN: group_unpack: Group G_samba did
> > not have any children
> > pengine[24820]: 2007/09/18_09:06:41 info: determine_online_status: Node
> > noemic1 is online
> > pengine[24820]: 2007/09/18_09:06:41 info: group_print: Resource Group:
> > G_mysql
> > pengine[24820]: 2007/09/18_09:06:41 info: native_print: R_192.168.12.197
> > (heartbeat::ocf:IPaddr2): Stopped
> > pengine[24820]: 2007/09/18_09:06:41 info: native_print: R_drbd_mysql
> > (heartbeat:drbddisk): Stopped
> > pengine[24820]: 2007/09/18_09:06:41 info: native_print: R_fs
> > (heartbeat::ocf:Filesystem): Stopped
> > pengine[24820]: 2007/09/18_09:06:41 info: native_print: R_mysql
> > (lsb:mysql): Stopped
> > pengine[24820]: 2007/09/18_09:06:41 info: group_print: Resource Group:
> > G_samba
> > pengine[24820]: 2007/09/18_09:06:41 notice: StartRsc: noemic1 Start
> > R_192.168.12.197
> > pengine[24820]: 2007/09/18_09:06:41 notice: StartRsc: noemic1 Start
> > R_drbd_mysql
> > pengine[24820]: 2007/09/18_09:06:41 notice: StartRsc: noemic1 Start R_fs
> > pengine[24820]: 2007/09/18_09:06:41 notice: StartRsc: noemic1 Start
> > R_mysql
> > crmd[24449]: 2007/09/18_09:06:41 WARN: Exiting pengine process 24820
> > killed by signal 11 [SIGSEGV - Segmentation violation].
> > crmd[24449]: 2007/09/18_09:06:41 ERROR: Exiting pengine process 24820
> > dumped core
> > crmd[24449]: 2007/09/18_09:06:41 info: crmdManagedChildDied: Process
> > pengine:[24820] exited (signal=11, exitcode=0)
> > crmd[24449]: 2007/09/18_09:06:41 ERROR: crmdManagedChildDied: The pengine
> > subsystem terminated unexpectedly
> > crmd[24449]: 2007/09/18_09:06:41 ERROR: do_log: [[FSA]] Input I_ERROR from
> > crmdManagedChildDied() received in state (S_POLICY_ENGINE)
> >
> > Are there any way to know what's the problem with pengine? Running
> > heartbeat_2.1.2-2_i386.deb for Ubuntu from Opensuse build service.
> >
> > Cheers
> >
> > Robert Lindgren
> >
> >
> _______________________________________________
> 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



More information about the Linux-HA mailing list