first time user with out of memory question

[Stephen Woodbridge]

timeless wrote:
> Ivan Krstic wrote:
>> Janning Vygen wrote:
>>   
>>> and stats slab produced "out of memory" message, too. At this time it was 
>>> impossible for many users to login to our web application.
>>>     
>> If memcached failure means your users can't log in, then you're using it
>> incorrectly.
>>   
> Janning is using the PHP Memcached session store. This is a fairly 
> common way to use Memcached. It does have the side effect that if a 
> session can't be created, the user can't log in.
> 
> I'm not convinced it's being used incorrectly. If a set() fails 
> regularly on out-of-memory in *ANY* usage, nasty things will occur.
>>> but i am concerned about it. As i have 3,5 GB dedicated to memached (-m 3584) 
>>> it should not run out of memory.
>>>     
>> I recommend compiling a new binary from the Facebook branch; it contains
>> a bunch of patches that generally make memcached saner.
>>   
> Apparently the Facebook patches include a fix to this very type of 
> problem, where even though you have more memory available to allocate, 
> you can't allocate any slabs of the size you need, and so get the "out 
> of memory" errors. Here we quote Steven Grimm:
> 
>     "I changed slabs.c so that it never gets those out-of-memory errors
>     (at the cost of exceeding the maximum memory setting if you fill up
>     your cache before you've allocated a slab of a particular class). "

When I was running my stress test, we gave memcached 1 GB of memory and 
it grew to 1.5 GB and then started throwing a lot of out of memory 
errors. I have not idea how the internals of this work, but based on 
some reading of the archives, I think this might be a symptom of memory 
fragmentation. But hey, I know not!

-Steve