I'm run top and I want to always see java/tomcat/redis/mysql/mongo, is that possible?
For most packages that require a yes/no answer, this works fine for me:
sudo apt-get install -y ....
But if you require a package that makes you accept a license, how do you automatically agree to that?
For example, sun java:
sudo apt-get install -y sun-java6-bin sun-java6-jdk sun-java6-jre sun-java6-plugin
This requires me to select OK, and then yes.
There are many companies that load a web application for you on their servers like say wordpress etc., and since they have other customers on the same server I was curious how they do this w/o having to restart the web server?
If they do restart the web server, it would take down other client sites and that wouldn't be a good situation to be in.
Is it possible with nginx?
I'm on ubuntu, is this the correct way to get puppet to start on reboots?
update-rc.d puppet default
If yes, where can I see the result of that command? Does it update a file that gets called on reboots?
Is this the same process for both the puppet master server and the puppet agent servers?
Update Just in case the puppet agent already runs after reboots using the default installs on ubuntu using:
sudo apt-get puppet facter
How can I know if is already setup to start on reboots?
From what I understand, LTS releases for servers are meant to be stable releases.
Now say I am running Ubuntu Lucid 10.04 LTS, if I get a particular package, like say puppet, it will load 0.25 or whatever.
Now if I want to load the latest version, what are my options?
I can get download via source and manually build, but then I have to setup start/stop scripts etc.
Is the only other option to update my source list? If I do that, what are the ramifications? Will it effect apt-get update/upgrade commands?
Does it even make sense, should I just go with the latest version like 11.04 Natty?
When you use sudo to perform root level operations, where does it log to by default on ubuntu?
I don't see entries in syslog.
I understand how puppet helps setup servers etc., but would you use puppet for something like:
apt-get update && apt-get upgrade
I get the feeling the answer is no, if that is the case, what sort of automated process would you suggest to do these kinds of tasks?
So say I am setting up a new instance (agent) to connect to puppet, how could I wrap together multiple commands that I want to run on the server just to prep it before connecting to the puppet master? I'm hoping I can just write out all the commands in a file, and then somehow run this file.
For things like:
- apt-get update && apt-get upgrade
- install ruby, other libs
- basic server lockdowns etc.
For a given url pattern, I want the contents of the file to be fetched from a CDN and not from my web server.
example:
www.example.com/some/path/ford/new?a=1&b=2
I need to extract:
ford new 1 2
I want to fetch the file from the s3 (or CDN) like:
example-com.s3.com/customers/ford.new.1.2.html
Is this possible at the HAProxy level or would I have to do this in nginx somehow?
The earlier in the pipeline the better obviously.
Possible Duplicate:
why do hosting co charge so much just for RAM?
Why is the cost of RAM so expensive from hosting companies? Just curious if anyone works at one and could explain it to me.
To get 16GB of RAM I've been quoted ~$700/month.
Is it the cost of energy? Is it because if you need 16GB you can afford $xxx/month?
Just getting my feet wet with puppet (just reading so far), and was hoping there were some sample (real life) manifests/recipes that setup a typical web server with things like apache/mysql/php etc. along with basic server hardening (ubuntu)
I have a folder that contains files for a static website like:
/site/index.html
/site/css/css.css
/site/js/js.js
/site/images/...
If I update something on my laptop, I want a single command to send the files off to my ubuntu server. I don't want to setup FTP on it if I don't have too, wondering if scp would be able to handle this?
Say my files on a Ubuntu server are like: /folder1/folder2/
I am in folder1 currently, and I want to move all files/folders (recursive) from folder2 to the current folder.
How can I do this?
I'm running nginx, and fastcgi to run a wordpress blog.
After a while of running, I am getting 502 bad gateway errors. If i reboot the server things work fine.
Other sites that are running on nginx are not effected (they are not php based), so I just figured out it must be the fastcgi process after looking at the log files saying the upstream client 127.0.0.1 is not responding.
How to I restart this? Is there a special log file for this fastcgi process?
I'm hoping there is a light weight, command line utility that I could use to setup a server from scratch.
I like python's fabric or ruby's capistrano but from what I was reading they are more used for deployment purposes and not setting up apache, mysql, update patches etc.
I know there are other tools like puppet, but I don't want to setup a master/slave for servers etc., I was hoping there was a more light-weight tool for this.
Can a single server run both haproxy and nginx at the same time?
I guess I would have to run nginx on a different port, and then route specific traffic to nginx?
Say I have 2 domains, 1 requires nginx and the other requires another service.
Can haproxy send requests for domain#1 to nginx? and domain#2 to another server?
I deleted the /var/log/nginx/error.log file, and then created a new one using:
sudo nano error.log
Doing ls -la shows that the error.log and access.log have the same permissions.
When I try and start nginx I get the error:
alert: could not open error log file: open() "/var/log/nginx/error.log" failed permission denited.
Update
When trying to start nginx, I am also seeing:
emerg: /var/run/nginx.pid failed 13: permission denied.
Is there a difference between a daemon and a service?
Or are they both basically an application that is resident in memory, and is bound to a specific port and listens/responds to requests?
Usually I re-request the CSR and install the certificate when I move servers.
Was hoping if there was a faster way, where I can just export/import a SSL certificate when I move my website to another server.
I am using Windows Server 2008 (R2).
On a Windows Server 2008, can you even telnet into the server? (not remote desktop, but via the command line)
If yes, is there much control really?
How do people automate if you can't telnet? (assuming you can't)
I was just thinking about how the platform choice effects costs in a SaaS type business.
If you were to choose between Linux and Windows, cost really seems to add up.
Please tell me if I am looking at things correctly.
Licenses wise, if you go with Windows, you will end up paying:
- USD$1K for a Windows 2008 Server license
- USD$10K for an SQL Server standard license.
Now for argument's sake let's say a single server can handle 1,000 users.
That would mean, you have a $11 cost per user for a year. Or you could break it out into maybe 2-3 years until you have to upgrade. So $3-4 /month cost per user.
On the Linux side, if you go with Java/PHP and MySQL, you have $0 license cost overhead.
It seems like this is a very simplistic way of looking at things, am I missing anything?
(Ignoring the productivity side of things, because that is developer specific.)