Hello List, I'm sure that my thought process is clouded right now and that I'm probably missing something obvious but here goes anyways...
Today I "upgraded" my working (but only half configured) 2.1.3 2-node cluster to heartbeat 2.99.x with pacemaker 1.0 from the SUSE repo referenced on the clusterlabs website. Both nodes are running on CentOS5, i386. After some initial configuration to get the packages installed (all normal dependency stuff), I went and started up the cluster. Once the cluster had started, I noticed that it had "rejected" (or otherwise not used) my previous cib and created a new, blank one instead. Using cibadmin, I tried to replace the current running cib with my old cib, but alas I get DTD errors. This is not unexpected as I understood that some things have changed between CRM and pacemaker. However, even when looking at the server logs, I don't get enough verbosity on the DTD violations to actually FIX the problems in my cib. I get the following type of information out from my logs: Oct 24 21:01:52 raza cibadmin: [17004]: info: Invoked: cibadmin -C -x newcib.xml -h raza.pet.ubc.ca Oct 24 21:01:52 raza cib: [16553]: ERROR: Element op failed to validate attributes Oct 24 21:01:52 raza cib: [16553]: ERROR: Expecting an element meta_attributes, got nothing Oct 24 21:01:52 raza cib: [16553]: ERROR: Invalid attribute id for element op Oct 24 21:01:52 raza cib: [16553]: ERROR: Element operations has extra content: op Oct 24 21:01:52 raza cib: [16553]: ERROR: Extra element operations in interleave Oct 24 21:01:52 raza cib: [16553]: ERROR: Element cib failed to validate content Oct 24 21:01:52 raza cib: [16553]: ERROR: cib_perform_op: Updated CIB does not validate against pacemaker-1.0 schema/dtd Oct 24 21:01:52 raza cib: [16553]: WARN: cib_diff_notify: Local-only Change (client:cibadmin, call: 2): 0.16.1 (Update does not conform to the configured schema/DTD) Without a line number to reference, stuff like "Element op failed to validate attributes" is pretty meaningless to me. I understand there is a problem with some attributes related to an operation but I don't get any information about what operation or what attributes of it did not validate. Do I just need to turn heartbeat debugging up or is there a better way to migrate a 2.1.x cib.xml to pacemaker 1.x? Maybe a tool that shows me what my DTD problems are, more specifically? I'm wondering what path others have taken during a similar upgrade and how they made their old cib compliant with the new DTD short of memorizing the whole DTD and doing all manual validation? On a side note, I also can't use the new hb_gui to configure (create or edit) any resources. Whenever I try to create a group or primitive, I get the following python barf: Traceback (most recent call last): File "/usr/bin/hb_gui", line 2822, in on_add objdlg = ObjectViewDlg(new_elem, True) File "/usr/bin/hb_gui", line 3900, in __init__ obj_view = ObjectView(self.xml_node, is_newobj, self.on_changed) File "/usr/bin/hb_gui", line 1994, in __init__ self.update(xml_node) File "/usr/bin/hb_gui", line 2130, in update self.on_after_show() File "/usr/bin/hb_gui", line 2138, in on_after_show self.obj_attrs.on_after_show() File "/usr/bin/hb_gui", line 6471, in on_after_show for widget in self.widgets[widget_type].values() : SystemError: Objects/funcobject.c:128: bad argument to internal function I'm using pacemaker-pygui-1.4-11.1.i386.rpm for the GUI package version. Any help would be appreciatd on either issue. As I said already, I really feel like I'm just missing something obvious here but it's late to be working and I might not be thinking straight. Thanks, --Ryan Thomson _______________________________________________ Linux-HA mailing list Linux-HA@lists.linux-ha.org http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems