[Spread-users] Spread Crash

[Rodrick Brown]

On Sat, Oct 4, 2008 at 12:00 PM, Adrian Revill <adrian.revill at shazamteam.com
> wrote:

> Hi,
>
> We are running spread 4.00.00 on a single segment with redhat el5, and have
> been experiencing a problem where one server crashes (OS), and all the
> spread daemons on all the other servers die.
>
> We have enabled logging on spread but do not see any reason logged,
> probably because spread uses a buffered logger and the buffer is not being
> flushed.
> We are also not getting any core dumps.
>
> Has anyone else seen this problem? or have any idea how to fault find?
>
> Also is there a newer version of spread?
>
> Our configuration is.
>
> Spread_Segment  172.20.255.255
> {
>       purple9         172.20.0.134
>       purple10        172.20.0.135
>       purple11        172.20.0.136
>       purple12        172.20.0.137
>       purple13        172.20.0.140
>       purple14        172.20.0.144
>       purple15        172.20.0.141
>       purple16        172.20.0.143
>       app11           172.20.0.130
>       app12           172.20.0.131
>       app13           172.20.0.142
>       wombat14        172.20.0.145
>       app15           172.20.0.132
>       app16           172.20.0.133
>       webportal11     172.20.0.138
>       webportal12     172.20.0.139
>       webportal13     172.20.0.157
>       app20           172.20.0.170
>       app21           172.20.0.171
>       purple20        172.20.0.172
>       purple21        172.20.0.173
>       app30           172.20.0.160
>       app31           172.20.0.161
>       purple30        172.20.0.162
>       purple31        172.20.0.163
>       pws30           172.20.0.164
>       pws31           172.20.0.165
> }
>
> DaemonUser = nobody
> DaemonGroup = nobody
> RuntimeDir = /usr/spread
>
> #       EXIT PRINT DEBUG DATA_LINK NETWORK PROTOCOL SESSION
> #       CONFIGURATION MEMBERSHIP FLOW_CONTROL STATUS EVENTS
> #       GROUPS MEMORY SKIPLIST ALL NONE
>
> #DebugFlags = { ALL !DATA_LINK !MEMORY !DEBUG !EVENTS }
> DebugFlags = { ALL !DATA_LINK !MEMORY !DEBUG !EVENTS }
>
>
Can you paste the last 10 or so lines in the log before the crash?  I've
seen weird memory behaviors with spread when you have members in too many
groups and sending out too many messages all at once where spread will
consume all memory on the host and crash with an malloc error.


>
>
> ______________________________________________________________________
> This email has been scanned by the MessageLabs Email Security System.
> For more information please visit http://www.messagelabs.com/email______________________________________________________________________
>
> _______________________________________________
> Spread-users mailing list
> Spread-users at lists.spread.org
> http://lists.spread.org/mailman/listinfo/spread-users
>



-- 
[ Rodrick R. Brown ]
http://www.rodrickbrown.com http://www.linkedin.com/in/rodrickbrown
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.spread.org/pipermail/spread-users/attachments/20081004/0fbc373f/attachment.html