It seems that I have an issue I can't resolve.
Every once in a while at seemingly the worst and most random times, the
service will stall out. It simply doesn't respond to anything.
Going into the computer and running htop shows that the computer is
doing nothing.
I am not running the server with a compiled version of my code.
I am running:
2.8.2-stable+timestamp.2013.11.28.13.54.07
(Running on Apache/2.2.22 (Ubuntu), Python 2.7.3)
The application does only basic uploading and loading of a single item.
Here is an example of the webplayer:
melodigram.com/melodigram/default/gram/b662a1da-8c2b-11e3-b0dc-1231390a0101
I can run F5 reloads on the site for 15 minutes straight, not have
anyone access the system then then try to load something, be it the
admin interface or the page I show you here and it site for a very long
time. The time it is unresponsive seems random, but that's because I
haven't discovered how to replicate the problem.
If I restart apache and try, it seems to work every time.
This issue is very annoying and I seem to hit this kind of thing in many
of the sites I am making and working on. We really need to figure out
what is going on.
I am running ubuntu 12.04 and set up the server using the scripts found
in web2py for postgres and so forth. I am using postgres and the issue
happens in both sqlite and postgres database setups.
I can't seem to replicate the issue when running locally with rocket
rather than apache.
The amazon server is a Medium instance. I thought maybe it was because
of the microinstance issues, but I seem to be wrong?
Any ideas into this?
Our app goes live next week and I really hope to resolve this issue.
The server is currently stalling out and has actually stopped mid page.
I have the htop running on the server on the right.
Last, is how do you think I could setup the server to better see what is
going on and solve the issue. It's really annoying knowing how simple
my service is, but seeing how very slow or inresponsive the site is.
BR,
Jason Brower