Re: ZODB eating CPU but RAM is free

Joni Orponen <[email protected]> Wed, 26 Sep 2018 16:12:52 +0200
Newsgroups gmane.comp.web.zope.zodb
Message-ID <CA+LW3i1UdPGqi8Yhucoto0ccJxDNJMOg7_Mb9togBP9wc=Wu0Q@mail.gmail.com>
On Wed, Jul 19, 2017 at 6:49 PM Jim Fulton <[email protected]> wrote:

> On Wed, Jul 19, 2017 at 6:53 AM, Joni Orponen <[email protected]>
> wrote:
>
>> On Wed, Jul 19, 2017 at 2:42 PM, Hanno Schlichting <[email protected]>
>> wrote:
>>
>>> Try increasing the zodb-cache-size for the zope2instance (client1). It
>>> defaults to 30000, try increasing it to 100000 and monitor the RAM usage.
>>>
>>
>> Is the cache_size_bytes option still considered to be experimental and/or
>> broken?
>>
>
> No. I don't consider it broken.
>

The estimates it produces are even more off than ballparking it via object
count. This could deserve a closer look, or do we differ on the semantics
of 'broken'? Or is there a known point onwards from which this behaves
better?

It is difficult to provision resources for a database the cache sizes of
which are unruly.

-- 
Joni Orponen

-- 
You received this message because you are subscribed to the Google Groups "zodb" 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.