Frequently Asked Questions

The title of the page serves as the heading to this section. It describes the problem that this page is solving.

In this page:

What is the difference between an Accelerator and a SmartMachine?

SmartMachine and Accelerator are both names used to describe the same thing – a zone running on Joyent SmartOS. Accelerator was the name used prior to May 2010.

What is my SmartMachine name?

Your SmartMachine name is the default name associated with your SmartMachine when delivered. It looks like: z12345AA, z12345AB, or fjsliyaa, fjsliyab. To find the name of your SmartMachine use the command

How can I find my SmartMachine public or private IP?

The public and private IP of your SmartMachine is in the welcome e-mail. Or you can find your public or private IP by using the command

Where can I get an SmartMachine image to use on my local computer?

Currently we do not have a publicly available SmarthMachine image. However, the SmartMachine is a pretty simple setup. We run Solaris Express Community Release and Pkgsrc.

• You can get the Solaris Express Community Release image at: http://opensolaris.org/os/downloads/sol_ex_dvd_1/

• See this list for our pkgsrc repository: http://wiki.joyent.com/accelerators:kb:pkgsrc:updating

How do I reset my SmartMachine to a clean image?

If you signed up for your SmartMachine through the MyJoyent control panel, delete the SmartMachine from your account and order a new SmartMachine as its replacement.

If you signed up for a SmartMachine another way, submit a ticket that you would like to wipe your SmartMachine clean and start fresh. Be sure you have everything backed up because we will wipe everything.

I can't reach Webmin. https://x.x.x.x:10000/ is not working?

If Webmin failing to load, you have one of two problems: The process is hung, or you are hitting your memory limit.

First, verify that you aren't hitting your memory limit. Log in to your SmartMachine, and use the prstat -Z command and look at the RSS column in the summary. If this value is near the size of your SmartMachine, shut down some services to free memory.

If this isn't the problem, restart the Webmin process:

My Path is not working ... shebang issues with /usr/bin/env

Truth be told, /usr/bin/env is just fine. The problem is really that 'env' was not originally intended as an all-cure to resolving binary paths automagically, and it will fail if it can't find e.g. ruby in its path. And the default system PATH does not include paths for the pkgsrc layer installed by us (/opt/local), which is typically/best deal with in interactive logins profiles (e.g. root's .profile).

To find out what your processes path is currently set to you do this (odds are you'll see it runs with a limited PATH):

for example looking for munin-node's process:

It's normally not a problem in your login shells, because those pull the full env from /root/.profile. It is a problem with e.g. SMF or Cron, because those run with a limited environment by default. So you want to do one of the following:

1. Always use full paths in scripts' shebangs;
2. Modify the system profile to include full PATH;
3. Change your SMF profile to set the right PATH.

#1 gets tedious especially with 3rd party scripts, and #2 is not highly recommended, so #3 is the best option I'd say. We do recommend that you always set PATH in SMF profiles.

Here is an example for mongrel setting the path:

Cron not running script

What is most likely happening here is two things:

1) path issue - cron runs under a different environment an it is best to use full paths to all programs in the script that is getting called.

For example if you are backin up MySQL with mysqldump you need to call that program with:

/opt/local/bin/mysqldump

2) permissions issue - make sure that the script is executable and the user's who cron you are using has permissions to:

• execute all programs in the script
• write to the directory (if backing up)
• access the database (if accessing a database)

Apache is in maintenance

Log rotation is set to happen at 0030 GMT, if you see that Apache has gone into maintenance at this time, then change the method of log rotation:

Fix Virtualmin settings

1. Log into Virtualmin
2. Follow this path (starting in the left pane):
   1. If you never upgraded Virtualmin: System Settings -> Module Config -> Log File Rotation
   2. If you upgraded Virtualmin and are runnign a fairly new version: System Settings -> Server Templates -> 'Default Settings' -> Log File Rotation
3. Change the Directives field to just '# postrotate' and save.
4. In the left pane, click 'Webmin' on top
5. Go to: System -> Log File Rotation -> Edit Global Options
6. Change the 'Truncate log file in place' option to 'Yes' and save.

Joyent provides Virtualmin as a third party opensource solution to help you manage your SmartMachine, but we are not responsible for the actual project. As such, we will make a best effort to solve your issue, however may need to turn to the Virtualmin Support community for help with it’s various features.

Fix existing domains

Check /opt/local/etc/logrotate.conf, and if there are any

kind of config blocks, remove the contents of the curly brackets, i.e.

Apache restarts at 0310

Check your settings in /etc/logadm.conf
Change this:

To this:

Add more SmartMachines to an existing account

You can add more SmartMachines directly in my.joyent.com. These machines are created without respect to any existing infrastructure you may have. If you have special requirements to ensure that your SmartMachine are set up to work well with your existing architecture you can contact Joyent Support at support@joyent.com with your request.

inetd service is in maintenance

• Looking at svcs -vx you see that inetd is in maintenance.
• Looking at the service log file you see these kinds of messages:
• Try to clear the service to see what happens:
• You then check with svcs -p inetd that it went back into maintenance.
• Then look at the output of {{dmesg}}You see a message like this:
• Then export that SMF file via
• Add where the rest of the propval lines are for inetd the line
• import the SMF file
• clear the inetd service
• check that it indeed came online:
• check that the dependent service is also online:

Unable to ssh/sftp due to encryption issues

In Ruby you will see something like this:

Or when trying to access via ssh from another application you may also see a cipher error – it means you need to add the cbc encryption types.

To add the cbc types of encryption do the following:

1. First change to root by enter the following from the command line:
2. You will be prompted for a password. Enter the password that is indicated in your welcome email for Root User Access
3. Now, run the following:
4. Then restart ssh for the changes to take effect:

And that's it, you should now be able to access via ssh or SFTP

Update Rmagick from RMagick-2.11.1 to RMagick-2.11.1nb1

If you are experiencing issues with RMagick, you can update to a new version easily:

You should be seeing 'ruby18-RMagick-2.11.1nb1' instead of just 'ruby18-RMagick-2.11.1'

disk usage (du) numbers not adding up

At times when you run du and compare it with the output of df the numbers will not add up.

This is most like caused by a file handle open to a file that has been deleted.

The command du won't count the file as existing, even though the filesystem still has that data allocated to a file.

When the process that's holding the filehandle open exits (this could be a database, or apache, or really any other daemon) the space consumption should go away and your filesystem usage should report correctly again.

The most typical thing that I've seen cause this is a log rotation that happens and nukes a log without properly HUP'ing the daemon.

finding large files filling up disk

To find what files and directories are filling up your disk you can use the command df in two ways:

(as root)

and follow the trail

The other way you can do this is with the command:

You can change the +200000 to be larger or smaller depending on how long of a list you are dealing with