Hi,
Please do not use huge amount of heap you will achieve nothing but GC stop
the world situation.
We do not use heap a lot, usually 512m or 1GB is good enough for server
installation the rest should put in disk cache.

On Sat, Jan 3, 2015 at 5:10 PM, <[email protected]> wrote:

> About second part of my question, it might sound stupid, I know. And I see
> now, that most of GC issues we had are gone (with inner limit, for
> example), but reason I was asking is just because of comments to this
> issue: https://github.com/orientechnologies/orientdb/issues/3318
>
>
> On Friday, January 2, 2015 2:47:58 PM UTC+5, [email protected] wrote:
>>
>> Dear OrientDB team, are there any rules of thumb on how big heap should
>> be depending on database size?
>>
>> Anyway, I suppose, that low memory should be cause to slower operations,
>> not to GC and OOM issues, am I right? (asking this because series of
>> repeated GCs almost always drove us to the "Timer already cancelled"
>> message and thus - to non-responsive engine)
>>
>> Another question - is there any relation between cache and heap memory
>> sizes. I mean - if we have 64GB of RAM - what is the best proportion
>> between heap and cache sizes should be, could we calculate it (there is a
>> chance, that heap memory should be bigger than some fraction of cache, or
>> vice versa).
>> What about 1GB available - is that scale, or have some special details to
>> consider?
>>
>> Thanks!
>>
>  --
>
> ---
> You received this message because you are subscribed to the Google Groups
> "OrientDB" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>



-- 
Best regards,
Andrey Lomakin.

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to