Hide Forgot
kexec-tools-2.0.0-273 contains automatic configuration of fence_kdump from Pacemaker configuration. Unfortunately, this support is tightly bound to Pacemaker and cannot be used for other clusters like oVirt/RHEVM. Proposed solution, which has already been merged to Fedora kexec-tools: - Add two new options to kdump.conf fence_kdump_nodes - List of hosts separated by space to send fence_kdump notification to (this option is mandatory to enable fence_kdump) fence_kdump_args - Command line arguments for fence_kdump_send (it can contain all valid arguments except hosts to send notification to) - Modify kdump behavior due to new options 1) If fence_kdump_nodes option is set and fence_kdump_send is found and executable -> configure network for kdump and execute fence_kdump_send with those nodes (and also with args specified in fence_kdump_args if not empty) 2) If fence_kdump_nodes is not set, try to configure fence_kdump using cluster settings (current behavior). This should stay in kexec-tools not to break compatibility and can be removed after Pacemaker will start using new options.
Hi, Martin Do you have a RHEL7 bug for this? We need to make it in RHEL7 before RHEL6... But it's too late for RHEL7.0 thus it should target for 7.1.. Thanks Dave
I cloned this bug for RHEL 7 as BZ1086988
Created attachment 886467 [details] [RHEL6.6 PATCH 1/4] Add function is_pcs_fence_kdump
Created attachment 886469 [details] [RHEL6.6 PATCH 2/4] Add function get_pcs_cluster_nodes
Created attachment 886470 [details] [RHEL6.6 PATCH 3/4] Code cleanup in fence_kdump setup section
Created attachment 886471 [details] [RHEL6.6 PATCH 4/4] Add fence_kdump support for generic clusters
Created attachment 886772 [details] [RHEL6.6 PATCH 1/4 v2] Add function is_pcs_fence_kdump
Created attachment 886773 [details] [RHEL6.6 PATCH 2/4 v2] Add function get_pcs_cluster_nodes
Created attachment 886774 [details] [RHEL6.6 PATCH 3/4 v2] Add function setup_cluster_nodes_and_options
Created attachment 886776 [details] [RHEL6.6 PATCH 4/4 v2] Add fence_kdump support for generic clusters
This feature is independent of any particular cluster software, everyone can use it as is. So I suggest this testing scenarios: 1) Test if no regressions were introduced with this patch to Pacemaker - Pacemaker uses its own method to configure fence_kdump support, it doesn't use new options fence_kdump_nodes and fence_kdump_args 2) Test if new options fence_kdump_nodes and fence_kdump_args works - only fence-agents-kdump RPM is needed to test this feature - I suggest this minimal scenario: - let's have host1 (it will contain fence_kdump configuration) and host2 (it will listen for fence_kdump messages) - On host1 add following option to kdump.conf (assuming that host2 is resolvable, otherwise IP of host2 has to be entered): fence_kdump_nodes host2 And restart kdump service - On host2 configure firewall to enable UDP connections to port 7410 - On host2 execute following command: /usr/sbin/fence_kdump -t 3600 -v 127.0.0.1 This will display all incoming fence_kdump messages (with above configuration every 10 seconds message should be received while host1 is kdumping) - Execute kernel crash on host1 using sysrq trigger echo c > /proc/sysrq-trigger
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHBA-2014-1502.html