What I mean is that nowadays it's not that expensive to have a 256GB server
with 20 cores. So instead of having a load balancer, a lot of machines
running web and application servers, a lot of other machines running
databases, etc. You can start with a small server and keep improving it for
@leonel, might i know what do you mean with scaling horizontally and
scaling vertically?
thanks and best regards,
stifan
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code.google.com/p/web2py/issues/list
I'm using a 32GB quad core server, which is way **way** more than I need at
the moment, but soyoustart is ridiculously cheap and I have this stupid
idea that before scaling horizontally you scale vertically as much as
possible.
--
Resources:
- http://web2py.com
- http://web2py.com/book
It seems that you are right, I set up a virtual machine with 4gb ram only
for the database and I ran the code along with redis on my laptop with 16gb
ram and my times now are nearly the 600ms. I think it is good enough.
Having some components that don't perform selects rendering only a simple
Yes I'm pretty sure it's a RAM issue, 48 free MB is way too low. You easily
start having to use virtual memory with a couple of requests.
To give you an idea, on a similar setup but a much better machine, I have a
much much more complex page than yours appears to be, with quite a lot of
Yes, the application, redis and the database are in the same server. This
server has 512 MB of RAM but I don't think I am using all the ram. I have
48mb free right now. Regarding the code, it just makes simple selects on
one or two tables (the querys arent doing anything weird). Do you still
That still seems slow actually. Is redis on that same machine or could you
have latency issues there too? If it is then it's hard to tell what's
causing the slowness now that everything is in the same machine. We would
need to know the machine specs and what your code is doing exactly. You may
Thank you to both of you. Anthony I was using redis to handle my sessions
hence I can discard that bottleneck besides I need the session info on my
components.
It seems that web2py performs more querys to the database than expected. I
have the database on digitalocean but was using web2py on
You are probably doing too much stuff. See if you can pre-compute stuff.
For instance. Imagine you have a debate table, then you have a comment
table that refers to the debate where the comment is made. If you want to
show a list of debates with the comment count, then you will have to make
9 matches
Mail list logo