On 09/08/2013, at 12:48 AM, RaSca <[email protected]> wrote:
> Il giorno Gio 08 Ago 2013 01:07:06 CEST, Andrew Beekhof ha scritto: >> On 08/08/2013, at 12:37 AM, RaSca <[email protected]> wrote: > [...] >>> The problem I got is that when I clone a VM (using virt-clone) >>> everything works fine until I try to add a new ping check. >> Can you more precisely describe what you mean by this? > > Of course: the steps for adding a new virtual machine are: > > - put the original resource in unmanaged; > - clone the original resource via virt-clone; > - add a primitive for the new vm; > - add an oredr/colocation constraint over the storage; > - add a location based upon the ping like this one: > > location loc_res_VirtualDomain_vm_connectivity res_VirtualDomain_vm \ > rule -inf: not_defined ping or ping lte 0 > > At this point somethings breaks up. The ping resource of the node where > the vm will be placed fails, making all the resources on it migrate. That sounds like something playing with the virt bridge when the vm starts. Is the host trying to ping through the bridge too? > >>> 1) Are there limitations about how many ping location can be declared? >> Well, there is a finite number of hosts that can be ping'd within a given >> interval. >> Is your timeout too short perhaps? Are you using fping which works in >> parallel? > > I'm not using fping (maybe this could be a solution) and the timeout of > the ping resource is 20 (which for me, makes sense). > >>> 2) Is this one (one vm = one ping location) the best practice to monitor >>> the connections of the nodes? >> ping resources were intended to check if a cluster node could reach the >> outside world. >> You're using them to check if a VM resource is alive? Perhaps David's >> remote-node stuff would be better suited. > > I'm using them to check if a resource (the vm) is on a node which can > reach the outside world. So one vm = one location. Is there a way to set > a location on an entire node so that if it looses the outside world all > the resources on it are migrated? I was convinced that this kind of > locations should be set up on the single resource. Many location constraints can reference the attribute created by a single ping resource. Its still not clear to me if you have one ping resource or one ping resource per vm... don't do the second one. > > Thanks, > > -- > RaSca > Mia Mamma Usa Linux: Niente รจ impossibile da capire, se lo spieghi bene! > [email protected] > http://www.miamammausalinux.org _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
