[Spread-users] Server dies

Jonathan Stanton jonathan at cnds.jhu.edu
Fri Jun 11 16:32:37 EDT 2004


Can you tell us what version of Spread you are running and on what type of 
platform (OS)? 

If the spread version is less then 3.17.2 could you try that first and see 
if it still has the problem? 

I have not seen this crash trigger before and it is one of 
those "Can't happen" cases that we put a check in for debugging. The 
'type' field that has value -12888 here should only ever have the value 0, 
1 or 2 so this looks like memory corruption. The fact that you can easily 
trigger it should make it easier to debug. 

If you are compiling spread yourself, would you mind making sure 
the "Makefile" has the "-g" option included on the CFLAGS= line and 
editing alarm.c and replacing the calls to exit(0) with "abort()" and then when it crashes the 
spread daemon should generate a core dump. Then email the core dump and 
the spread binary itself to me personally (it will be too large for the 
list) and I'll look at it.



On Fri, Jun 11, 2004 at 10:47:23AM +0000, G Hasse wrote:
> Hello all.
> We very often gets a 
> ----
> Smallest_rep: bug! current index 0 is proc -1073754200 of type -12888
> Exit caused by Alarm(EXIT)
> ----
> and the server dies.  I know that there is problems with
> "small messages" but why does the server die? 
> We cant run the server for more than some minutes, and that
> doesent feel reliable...
> -- 
> Göran Hasse
> ----------------------------------------------------------------
> Göran Hasse            email: gh at raditex.se     Tel: 08-6949270
> Raditex AB             http://www.raditex.se    Fax: 08-4420570
> Sickla Alle 7, 1tr                              Mob: 070-5530148
> 131 34  NACKA, SWEDEN
> _______________________________________________
> Spread-users mailing list
> Spread-users at lists.spread.org
> http://lists.spread.org/mailman/listinfo/spread-users

Jonathan R. Stanton         jonathan at cs.jhu.edu
Dept. of Computer Science   
Johns Hopkins University    

More information about the Spread-users mailing list