[Doc] Minimum hardware requirements to run MAAS.

[Jeroen T. Vermeulen]

I set up a 1.7 beta4 instance (version 1.7.0~beta4+bzr3130-0ubuntu1~trusty1 to be exact) and started an import of Ubuntu 14.04 images for amd64. When it finished a few hours later, the Images page in the UI said:

«Boot image import process not started. Nodes will not be able to provision without boot images. Start the boot images import process to resolve this issue.»

and

«No Ubuntu images have been imported. Ubuntu images are required to allow nodes to enlist, commission, and install.»

The maas.log has a lot of messages like:

«
Sep 30 11:50:41 test-1 maas.import-images: [INFO] Started importing boot images.
Sep 30 11:50:41 test-1 maas.import-images: [WARNING] No resources found in Simplestreams repository u'http://192.168.122.74/MAAS/images-stream/streams/v1/index.json'. Is it correctly configured?
Sep 30 11:50:41 test-1 maas.import-images: [WARNING] Finished importing boot images, no boot images available.
Sep 30 11:52:29 test-1 maas.tftp: [WARNING] No boot images have been imported yet.
Sep 30 11:55:29 test-1 maas.tftp: message repeated 9 times: [ [WARNING] No boot images have been imported yet.]
»

[Blake Rouse (blake-rouse)]

Check the maas.log for maas.bootresources. maas.import-images is on the cluster, and the error you are seeing is because your region has no images. Be sure that maas.bootresources is showing in the log, if not. Be sure to start the import process from the image page on the UI, or over the API: maas admin boot-resources import.

[Jeroen T. Vermeulen (jtv)]

I did start the import process from the UI. In maas.log I see both starts and completions of maas.bootresources, though a lot more starts than completions. The completions always seem to come a few seconds after start.

[Blake Rouse (blake-rouse)]

Provide me the output of the following commands:

maas admin boot-sources read
maas admin boot-resources read
maas admin boot-images read cluster_uuid

[Jeroen T. Vermeulen (jtv)]

The VM where this originally happened no longer exists. But I set up a new one with a fresh package built from today's trunk, and ran the imports to completion. The problem persists.

Those outputs you asked for:

$ maas admin boot-sources read
[
    {
        "url": "http://maas.ubuntu.com/images/ephemeral-v2/releases/",
        "keyring_data": "",
        "resource_uri": "/MAAS/api/1.0/boot-sources/1/",
        "keyring_filename": "/usr/share/keyrings/ubuntu-cloudimage-keyring.gpg",
        "id": 1
    }
]

$ maas admin boot-resources read
[]

$ maas admin boot-images read 683726de-c393-40ce-adc1-5e2a892ca20d
[]

[Jeroen T. Vermeulen (jtv)]

This may also come in handy:

$ sudo ls -alR /var/lib/maas/
/var/lib/maas/:
total 20
drwxr-xr-x 5 maas maas 4096 Oct 3 02:37 .
drwxr-xr-x 45 root root 4096 Oct 3 02:36 ..
drwxr-xr-x 2 maas maas 4096 Oct 3 02:29 boot-resources
-rw-r--r-- 1 root root 0 Oct 3 02:37 dhcpd6-interfaces
-rw-r--r-- 1 root root 0 Oct 3 02:37 dhcpd-interfaces
drwxr-xr-x 2 maas maas 4096 Oct 3 04:37 gnupg
drwxr-xr-x 2 maas maas 4096 Oct 3 02:37 media

/var/lib/maas/boot-resources:
total 8
drwxr-xr-x 2 maas maas 4096 Oct 3 02:29 .
drwxr-xr-x 5 maas maas 4096 Oct 3 02:37 ..

/var/lib/maas/gnupg:
total 12
drwxr-xr-x 2 maas maas 4096 Oct 3 04:37 .
drwxr-xr-x 5 maas maas 4096 Oct 3 02:37 ..
-rw------- 1 maas maas 0 Oct 3 02:37 pubring.gpg
-rw------- 1 maas maas 1200 Oct 3 02:37 trustdb.gpg

/var/lib/maas/media:
total 8
drwxr-xr-x 2 maas maas 4096 Oct 3 02:37 .
drwxr-xr-x 5 maas maas 4096 Oct 3 02:37 ..

[Blake Rouse (blake-rouse)]

Please upload the mas.log and mass-django.log as I am unable to reproduce this issue. Hopefully somewhere in that log we can see why the region is not importing images.

[Jeroen T. Vermeulen (jtv)]

I think I see the problem in the maas-django.log... But it's a weird one!

[Jeroen T. Vermeulen (jtv)]

Not much in maas.log though.

[Jeroen T. Vermeulen (jtv)]

I think I know! The VM ran out of memory. Linux's OOM killer has a grudge against postgres. And it ran several times. Getting killed and restarted would kick the database server into recovery mode.

Now, this is by no means a beefy VM. But it's got half a GB of memory. With no nodes, one user, and images to import for just one Ubuntu release on one architecture, why does it run out of memory? Surely it's not the file upload thing again?

[Jeroen T. Vermeulen (jtv)]

I set up ‘swapspace’ on the VM, and then the import worked!

(Shameless plug: Swapspace is a dynamic swapfile manager. Just install it and you'll notice no difference... until you start running low on memory and it magically starts increasing your swap space. It's robust, lightweight, and autonomous.)

[Blake Rouse (blake-rouse)]

Since the import images are stored in the database more memory is needed. We should come up with minimum hardware specs to run MAAS, to fix others that will have this issue.

[Andres Rodriguez (andreserl)]

Dear user,

This is an automated message.

We believe this bug report is no longer an issue in the latest version of MAAS. For such reason, we are making this issue as Won't Fix. If you believe this issue is still present in the latest version of MAAS, please re-open this bug report.