[Linux-HA] Re: mgmtd problem on suse 10.3

[joe]

Lars Marowsky-Bree wrote:
> On 2007-10-10T22:18:35, joe <joe at tmsusa.com> wrote:
> 
>> Any hints or words of wisdom would be welcome at this point.
> 
> Use the packages from
> http://download.opensuse.org/repositories/server:/ha-clustering/

I was hoping that would be a quick fix, but no joy. I've installed the
heartbeat packages from:

http://download.opensuse.org/repositories/server:/ha-clustering/

but hb_gui shows the same symptoms as with my self compiled version.

Here is the syslog output from my hb_gui login:

Oct 11 20:32:42 ahl mgmtd: [27104]: debug: recv msg: login hacluster **** 1.0
Oct 11 20:32:42 ahl mgmtd: [27104]: debug: send msg: ok
Oct 11 20:32:42 ahl mgmtd: [27104]: debug: recv msg: regevt evt:cib_changed
Oct 11 20:32:42 ahl mgmtd: [27104]: debug: send msg: ok
Oct 11 20:32:42 ahl mgmtd: [27104]: debug: recv msg: regevt evt:disconnected
Oct 11 20:32:42 ahl mgmtd: [27104]: debug: send msg: ok
Oct 11 20:32:42 ahl mgmtd: [27104]: debug: recv msg: all_nodes
Oct 11 20:32:42 ahl mgmtd: [27104]: ERROR: Cannot start node walk
Oct 11 20:32:42 ahl mgmtd: [27104]: ERROR: REASON: can't send message to IPC
Channel: Resource temporarily unavailable
Oct 11 20:32:42 ahl mgmtd: [27104]: debug: send msg: fail


Someone suggested doing an strace of mgmtd, how exactly would I do that, an
"strace mgmtd" entry in ha.cf?

I guess my idea at this point is to try and write some test programs to
exercise the ipc functions and try to isolate it a little better.

Any suse folks out there successfully running mgmtd on 10.3?

Joe