A lot of my Innodb tables require 'defragmentaion'. My understanding is that 'OPTIMIZE' may not be the best way to do it. Some people suggest doing a 'ALTER TABLE' to have Mysql rebuild (and therefore defragment) the table.
Does anyone have any suggestions (based on experience) of whats the best way to defragment production Innodb tables?
On my local dev machine (Ubuntu 10.10) a web application seems to consume approximately 10Mb of memory (according to top). This is Apache2, pre fork, php5.3
On my server which is the same basic setup as my local dev machine, apache is consuming approximately 30Mb per processes (via top).
I cant think of any modules that production has that my local copy doesn't (php or apache). The only difference is that production is a lot busier than local (obviously) but I didn't think that would really make such a difference.
My question is, what can I do to troubleshoot this and work out why apache2/php is using so much ram on production per apache process?
I have a wildcard SSL certificate setup on *.domain.com and its working fine. I now want to add a single ssl certificate to domain.com.
Is this likely to cause me problems? I assume not, but I thought it would be good to check that someone else has done it.
I have a application that receives data from approximately 1000 different sources every couple of minutes. This data needs to be saved into a OSS database (most likely MySQL - but depending on answers I'd consider switching if appropriate).
The 1000 external servers that are sending the information will often request data every 2 minutes from the application.
The application is data critical in that it cannot afford to ever be down. I have the application part sorted for HA (High Availability), but I'm after suggestions for database HA. Oracle is not an option.
A friend suggested SQLlite rsynced between the application servers but that sounds dangerous to me. I looked into a MASTER-MASTER MySQL db setup, but it looks problematic and from users comments potentially unstable.
Any suggestions?
Has to run on linux, has to be open source.
I am about to setup a web application that consists of 2 app servers (running php/apache) and 2 db servers (ubuntu slices on slicehost).
The application will be sending approximately 5000 emails a day initially which will grow to 20k-100k within about 3 months as new features get rolled out.
I have never really administered a mail server before so I am not sure what kind of load will start impacting performance.
The mail server (postfix) will have spikes where it will go from 5 emails a minute, to 20k a minute. (although this might be staggered, lets assume its not staggered for now)
Given the basic outline above, do you think it would be beneficial to run another server (slice) which is purely a mail server (outbound, running postfix) thereby shifting all email load over to that server instead of the app servers. Or, is postfix efficient enough that there is no need and a separate server is possibly overkill?
If people could refrain from "Test it and see" that would be good, I am asking for people with experience to comment.
Additional note: Postfix will only be sending email, receiving email will be handled by google apps.
I would like to implement dynamic sub domains for a site that has to be secure (ssl).
I can setup a call to a dns api to add the A name for the subdomain, but was wondering if anyone can recommend an approach for adding the ssl information (based on a wildcard ssl certificate)?
All the subdomains would be for the name main domain.
eg.
The webserver is apache serving php.
I know these questions are hard to answer :) But I want to get some kind of an idea of the impact of separating services onto different servers.
Specifically, I am setting up a website that will have a media server (apache serving static files), a app server (apache processing php files), a master db server and a slave db server.
I was going to run memcached on both the media and app servers as a shared pool, but was wondering if it was expensive (in terms of time and resources) to do so? Obviously everytime a memcached call is requested (from the app server) it has to make a tcp connection to the media server, work out where the results are and then return them.
For a small website setup like this, would it be better to bump up the ram on the app server and NOT pool memcached across servers? Or is the difference so small its not worth worrying about?
Even though I used memcached as the example, I would prefer if answers are kept fairly generic as the same scenario could be applied to other services (like a db).
To be more specific, I plan to setup a load balanced e-commerce site. The two load balancers will accept https connections from the tubes, then as is the nature of ssl, they will need to verify the connection, decrypt it, then forward the UN-encrypted request onto the application servers.
In a secure private network that would all be well and good, however I want to run this setup with virtual machines (slicehost). Theoretically the machines should be on a private network, however it is feasible that another persons machine (slice) could somehow snoop on traffic within the network. I need to assume that the network is not secure.
So. I have thought about using ssh and someone suggested to me to use a VPN (which I have not setup before and don't really know a lot about). If I use ssh I would need to be continually monitoring the connection and making sure its up.
I guess I'd like to know what other people do?
The less complicated the better (ie. the less things that can go wrong and need monitoring)
This is a linux setup. I'm thinking about using pound for the load balancer. Still need to do some tests on that side of things. Have started to read up on nginx as well.