Re: Scalability in postgres

On Thu, 4 Jun 2009, Mark Mielke wrote:

> david(at)lang(dot)hm wrote:
>> On Thu, 4 Jun 2009, Mark Mielke wrote:

>>> An alternative approach might be: 1) Idle processes not currently running
>>> a transaction do not need to be consulted for their snapshot (and other
>>> related expenses) - if they are idle for a period of time, they
>>> "unregister" from the actively used processes list - if they become active
>>> again, they "register" in the actively used process list,
>> how expensive is this register/unregister process? if it's cheap enough do
>> it all the time and avoid the complexity of having another config option to
>> tweak.
>
> Not really relevant if you look at the "idle for a period of time". An active
> process would not unregister/register. An inactive process, though, after it
> is not in a commit, and after it hits some time that is many times more than
> the cost of unregister + register, would free up other processes from having
> to take this process into account, allowing for better scaling. For example,
> let's say it doesn't unregister itself for 5 seconds.

to do this you need to have the process set an alarm to wake it up. if
instead it just checks "anything else for me to do?, no?, ok I'll go
inactive until something comes in" you have a much simpler application.

the time needed to change it's status from active to inactive should be
_extremely_ small, even a tenth of a second should be a few orders of
magnatude longer than the time it needs to change it's status.

this does have potential for thrashing if you have lots of short delays on
a lot of threads, but you would need to have multiple CPUs changing status
at the same time.

>>> and 2) Processes could be reusable across different connections - they
>>> could stick around for a period after disconnect, and make themselves
>>> available again to serve the next connection.
>> depending on what criteria you have for the re-use, this could be a
>> significant win (if you manage to re-use the per process cache much. but
>> this is far more complex.
>
> Does it need to be? From a naive perspective - what's the benefit of a
> PostgreSQL process dying, and a new connection getting a new PostgreSQL
> process? I suppose bugs in PostgreSQL don't have the opportunity to affect
> later connections, but overall, this seems like an unnecessary cost. I was
> thinking of either: 1) The Apache model, where a PostreSQL process waits on
> accept(), or 2) When the PostgreSQL process is done, it does connection
> cleanup and then it waits for a file descriptor to be transferred to it
> through IPC and just starts over using it. Too hand wavy? :-)

if the contents of the cache are significantly different for different
processes (say you are servicing queries to different databases), sending
the new request to a process that has the correct hot cache could result
is a very significant speed up compared to the simple 'first available'
approach.

>>> Still heavy-weight in terms of memory utilization, but cheap in terms of
>>> other impacts. Without the cost of connection "pooling" in the sense of
>>> requests always being indirect through a proxy of some sort.
>> it would seem to me that the cost of making the extra hop through the
>> external pooler would be significantly more than the overhead of idle
>> processes marking themselvs as such so that they don't get consulted for
>> MVCC decisions
>
> They're separate ideas to be considered separately on the complexity vs
> benefit merit.
>
> For the first - I think we already have an "external pooler", in the sense of
> the master process which forks to manage a connection, so it already involves
> a possible context switch to transfer control. I guess the question is
> whether or not we can do better than fork(). In multi-threaded programs, it's
> definitely possible to outdo fork using thread pools. Does the same remain
> true of a multi-process program that communicates using IPC? I'm not
> completely sure, although I believe Apache does achieve this by having the
> working processes do accept() rather than some master process that spawns off
> new processes on each connection. Apache re-uses the process.

the current limits of postgres are nowhere close to being the limits of
fork.

on my firewalls I use a forking proxy (each new connection forks a new
process to handle that connection), I can get connection rates of tens of
thousands per second on linux (other systems may be slower).

but the here isn't the cost of establishing a new connection, it's the
cost of having idle connections in the system.

David Lang