On Tue, 07/03/2012 01:16 AM, Errol Neal <eneal@businessgrade.com> wrote:
> Hi All. I've been looking for examples of how to use the portblock RA and I'm not really finding what I need.
> <SNIP>
> Jul  2 20:32:11 nas1 lrmd: [15349]: info: RA output: (unblock1:monitor:stderr) 0 bytes (0 B) copied
> Jul  2 20:32:11 nas1 lrmd: [15349]: info: RA output: (unblock1:monitor:stderr) , 0.019563 s, 0.0 kB/s
> Jul  2 20:32:38 nas1 cib: [15347]: info: cib_stats: Processed 125 operations (2320.00us average, 0% utilization) in the last 10min
> Jul  2 20:32:41 nas1 lrmd: [15349]: WARN: unblock1:monitor process (PID 28154) timed out (try 1).  Killing with signal SIGTERM (15).
> <SNIP>
> 
> I haven't dug into the portblock RA script yet, but I can see that the command line for PID 28154  is:
> 
> dd of=/srv/samba/shares/data/tickle/172.24.100.15.new conv=fsync
> 
> The tickle dir is an OCFS2 volume. 
> 
> It seems that after that process starts, then the portblock resource fails and attempts to stop samba and ctdb. 
> 
> Any thoughts?
> 

I think I may have solved my own issue. From my perspective, it appears that the portblock resource behaves much better on local storage. I set the tickle_dir to /dev/shm and the sync_script to csync2 and now things are behaving themselves - though OCFS2 is still hanging/crashing after simulating a failure (e.g power loss, sudden reboot). That is all all together different matter for a different list though.