max-cache-size doesn't work with 9.5.0b1

Attila Nagy bra at fsn.hu
Mon Jan 28 16:10:28 UTC 2008


Hello,

On 01/23/08 21:07, JINMEI Tatuya / 神明達哉 wrote:
> (Catching up an old issue....)
>
> At Sat, 08 Dec 2007 22:40:14 +0100,
> Attila Nagy <bra at fsn.hu> wrote:
>
>   
>> I have a somewhat loaded caching only nameserver, running FreeBSD/amd64 7.0.
>>
>> I decided to try 9.5.0 just to see how well it goes, but it turned out 
>> that its memory usage grows endlessly (OK, limited by the total VM space :).
>>
>> I see this in top (the usage grows nearly every seconds, the machine 
>> gets some thousands queries per sec):
>>   PID USERNAME  THR PRI NICE   SIZE    RES STATE  C   TIME   WCPU COMMAND
>> 16969 bind        5  96    0   727M   723M select 0   8:51 68.12% named.950
>>
>> The relevant parts from the config (cache-size set to very low, just to 
>> demonstrate the problem):
>>         max-cache-size 32M;
>>         cleaning-interval 15;
>>
>> This is the case with and without ISC_USE_INTERNAL_MALLOC.
>>
>> With 9.4.2 the memory usage does not grow above a given size:
>> 17727 bind        5  97    0   131M   111M select 0  11:14 11.08% named.942
>>     
>
> If you have time, could you rebuild named as follows
>
> % STD_CDEFINES='-DLRU_DEBUG' ./configure; make
>
> and try again?  This won't solve the problem, but provide more
> detailed log messages tracking the memory usage of the cache.  Make
> sure threads are disabled (this debug feature is currently
> thread-unsafe and trigger an assertion failure when used with multiple
> threads).
>   
Without threading I don't see this effect, the memory usage stops at a 
sane limit and it's size can be affected by setting the max-cache-size 
option.

I don't think you would gain anything usable with that, am I right?

Thanks,



More information about the bind-users mailing list