Hal Murray wrote:
>> Has anybody else upgraded a pool server to the rc1 version of ntpd?
>> Id did and my stratum one box has died twice with out of memory in the
>> past three days.    I'm wondering if the load of the pool clients is
>> triggering a bug .... 
>>     
>
> How long does it run before it dies?  There is some memory allocated for each 
> of the last N systems that have talked to your system.  It has a max of 600.  
> How long before you hit that max?  I'd expect memory usage to stop growing 
> when it gets that far.
>
> Can you track how much memory it is using over time?  and/or how many systems 
> have used your server?  Can you switch back to the old system and get similar 
> info to compare with?
>
> Are you using anything fancy like crypto?
>
> What OS?  ...
>
> It might just be that the new version needs slightly more memory and you were 
> right on the edge.
>
>
>
>   
FreeBSD5.[34] - I moved it to an internal server to eliminate the load -
the copy I started around noon (at around 2.5mb resident) is now 7.8mb
resident some 7 hours later.  I downgraded the S1 server to the
stable-20060224 build and it's steady at about 3.5 mb (about 200k more
than it started with).  See my post to c.p.t.ntp withe full ntp.conf - 
looks like there is a memory leak in rc1.  I'd love to have somebody
else confirm it. 

John

_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers

Reply via email to