Kyle Cureau's questions

Typically I set ssh aliases as follows:

ssh -i ~/path/key.pem ubuntu@public-host-name.compute-x.amazonaws.com

But with EC2 instances, the public-host-name changes whenever an instance is stopped and restarted.

My EC2 instances are on a VPN and have constant private IPs. What strategy can I use so that I can log into my instances without constantly changing the alias?

Thanks!

A MySQL cloud server only allows remote access to a certain server (in this example, 11.22.33.44). Running the following commands from the terminal is verified to work:

1. ssh -p 40000 user@11.22.33.44
2. mysql -u dbuser -p -h example.com

But when combining them into command (in order to assign it to an alias), it doesn't work:

ssh -p 40000 user@11.22.33.44 "mysql -u dbuser -p -h example.com"

No error message is displayed, and the terminal never outputs a returned response.

Any ideas?

I'm running XAMPP on Windows 7. error_log("test") doesn't output to the my error_log specified in my vhost config (although the access log specified in the same config) does work.

Any ideas?

<VirtualHost *:80>
  DocumentRoot "c:\Users\User\Example"
  ServerName example.local

  CustomLog "C:\Users\User\Program\logs\local_access.log" combined
  ErrorLog "C:\Users\User\Program\logs\local_error.log"

  <Directory "C:\Users\User\Program">
    Options Indexes FollowSymLinks
    AllowOverride All
    Order allow,deny
    Allow from all
  </Directory>
</VirtualHost>

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
19857 mysql     20   0  514m  47m 3404 S  386  9.7   4511:45 mysqld
15659 www-data  20   0  205m  11m 4856 S    7  2.3   0:00.67 apache2
15701 www-data  20   0  205m  11m 4424 S    5  2.2   0:00.19 apache2
15700 www-data  20   0     0    0    0 Z    1  0.0   0:00.02 apache2 <defunct>

These are my results from top on Ubuntu. It was curious as to why my site keeps going down today. Could this be it?

Why is MySQL using 386% CPU? Is this normal? What can I do to fix it?

I have a deploy script that git pulls code from a repo. I'd like to minify the javascript (js) and css files in that deploy script using yuicompressor.

In order to minify one js file you'd use something like this:

java -jar yuicompressor-2.4.2.jar test.js -o test.js

How can I do that to all js and css files recursively (inside subdirectories and across all files while excluding files that aren't js or css)?

Thanks!!

I have a hourly cronjob script that curls a php script. The php script isn't being executed. In crontab -e I have:

1 * * * * /var/www/path/to/script.sh

Doing grep cron /var/log/syslog outputs:

Mar 23 21:45:01 Emile CRON[5339]: (root) CMD (/var/www/path/to/anotherscript.sh)
Mar 23 22:01:01 Emile CRON[5386]: (root) CMD (/var/www/path/to/anotherscript.sh)
Mar 23 22:01:01 Emile CRON[5387]: (root) CMD (/var/www/path/to/script.sh)
Mar 23 22:14:01 Emile CRON[5425]: (root) CMD (   cd / && run-parts --report /etc/cron.hourly)
Mar 23 22:15:01 Emile CRON[5429]: (root) CMD (/var/www/path/to/anotherscript.sh)

I've verified that the url that is being curl in script.sh is correct and that the path in crontab -e is correct. I also verified that the php script works when manually executed. What else is there? Any ideas? Another way to diagnose? What's that cd / && run-parts in my syslog?

Thanks in advance for the help!

Whoa, acronyms :) So according to this question you can have multiple SSL subdomains on one IP address as long as your server supports TLS (which Apache 2.2x does).

Another answer to that question points out that the client browser must have SNI support to work, which IE doesn't have on Windows XP. So, what happens to people with that browser is a warning message saying that the SSL cert doesn't match the domain.

Can one resolve this issue for those client browsers without SNI support? Does a wildcard (for subdomains) do the trick? Are there other (cheaper) options?

I just checked my server's /var/log/auth.log and found that I'm getting over 500 failed password/break-in attempt notifications per day! My site is small, and its URL is obscure. Is this normal? Should I be taking any measures?

My website has an SSL cert (example url: https://subdomain.example.com). Under Apache it's set up for both port 80 and port 443. So under the following configuration, anyone who goes to http://subdomain.example.com is sent to https://subdomain.example.com . But for visits from Internet Explorer, the redirect doesn't happen. Instead, http visits get a "Internet Explorer cannot display the web page." with a list of client-side solutions to try.

Any ideas on how to fix the config so IE visits have the same behavior as the other browsers (that is, send http to https automatically)?

NameVirtualHost *:443

<VirtualHost *:80>
  DocumentRoot /var/www/somewebroot
  ServerName subdomain.example.com
</VirtualHost>

<VirtualHost *:443>
     DocumentRoot /var/www/somewebroot
     ServerName subdomain.example.com
    #   SSL CERTS HERE
</VirtualHost>

*Tested IE8, IE9 beta