> Tribblix on AWS

AWS - Amazon Web Services - is one place to run Tribblix in the cloud.

You can run Tribblix on an EC2 instance on AWS.

There are AMIs available for both vanilla Tribblix and the LX-enabled OmniTribblix. These are available in the London (eu-west-2) region. If you set your region to London, launch an instance, browse AMIs, and then search Community AMIs for "illumos" or "Tribblix" then you should get some results like the following:

Specifically, the currently available and supported AMIs, which can be used for both Nitro and non-Nitro instances, are:

• Tribblix m35 - ami-01b06698db06cdb16
• OmniTribblix m35lx - ami-09cae0ad1eebc2b0e

The previous release of supported AMIs are:

• Tribblix m33p1 - ami-03745614c82800523
• OmniTribblix m33p1lx - ami-0608ad0511e6fe3e1
• Tribblix m33p1 (Nitro) - ami-0941a5d7b466ea8fd
• OmniTribblix m33p1lx (Nitro) - ami-0554b8df5122ed8c7

If you want to run Tribblix in a different region, then you'll have to copy the AMI.

Sometimes the networking doesn't initialise properly on the first boot. Rebooting the instance from the AWS console ought to clear that.

For releases prior to m35, the images labelled (Nitro) support the new (t3, m5 etc.) instance types. They won't run on the older (t2, m4, etc.) instances. Likewise, the images without the (Nitro) label will only run on the older instance types. As of the m35 release, Nitro support has greatly stabilised and the same images can be used for either.

If you see Time of Day clock errors on the console, which will manifest itself as silly clock speeds reported by psrinfo -v, or sleep taking a wildly different time to what you asked for, then

echo set pit_is_broken = 1 > /etc/system.d/pit

and reboot, and all should be fine.

The m4 and similar generation instances work well, as do t2 instances. The t2.micro is just fine for Tribblix, and is on the free tier. With the current AMIs, the newer Nitro instances are a good bet, and cheaper.

I recommend you enable static networking (again, not on Nitro instance with older AMIs) with

zap staticnet -y

After launching an instance, you should be able to log in as root with your key pair. That's the only way in, there's no jack account, and no valid root password. You should create your own regular users, of course.

If you want to store data, add an EBS volume and create a separate ZFS pool. (This will also make it easier to move the data to another instance when you need to.)

The images supplied only have a smallish system disk. If you need a bigger root pool (as opposed to adding a second data pool) then create the root EBS volume with a larger size and expand the pool with:

# zpool online -e rpool c2t0d0

to get the additional space recognised. You can enlarge an EBS volume after instance creation, but it will take a reboot for Tribblix to realise that the volume size has changed.