On Tue, Oct 26, 2010 at 09:47:25AM -0700, dan clark wrote: > Hello Gentle Readers! > > Several folks have talked about the advantages of the well documented > Availability Management Framework (AMF) promoted by the Service Availability > Forum (SAF). With the strength of the fundamental messaging system of > corosync one might think that the use of openais to provide such a framework > would be a good fit. As noted earlier on the list there are competing > capabilities (such as heartbeat on top of corosync and the SAM interface > within corosync). Is it possible to experiment with a simple AMF > implementation under openais at this point, even with the noted limitations > of 1/n failover (versus nxm)? > > Under the debugger the corosync daemon never seemed to call > 'amf_lib_init_fn' or amf_sync_init' associated with the amf_service_engine. > > > Is it the case that corosync/exec/evil.c evil_callbacks_load is actually > disabling the startup of the openais AMF implementation, or is this just an > extra set of callbacks that will eventually be removed across all > frameworks? > > With logging turned up it appeared that the AMF implementation initialized > once setup in /etc/corosync.conf > > corosync [SERV ] service.c:corosync_service_link_and_init:265 Service > engine loaded: openais availability management framework B.01.01 > corosync [SYNC ] sync.c:sync_deliver_fn:451 Synchronization actions > starting for (dummy AMF service) > > There was, however, no indication of the parsing of amf.conf. > > Is it possible to log the parsing of amf.conf to see how it is being > processed? > > Respectfully, > dan
The AMF service in OpenAIS is not functional. One could make the argument that it should not be included releases. Ryan _______________________________________________ Openais mailing list [email protected] https://lists.linux-foundation.org/mailman/listinfo/openais
