Thread Info
  • State Verified Answer
  • +1 person also asked this people also asked this
  • Locked Locked
  • Replies 9 replies
  • Subscribers 8 subscribers
  • Views 2617 views
  • Users 0 members are here
Options
Suggested
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Sophos UTM AMIs not starting with AWS instance type m3.medium

Martin Garbe

We are using Sophos UTM in AWS with PAYG license. Some month ago we changed the instance type from t2.small to m3.medium. This worked fine. We live in eu-west-1.

Now setting up another instance with the same AMI [1] and same instance type (m3.medium) fails on boot.

Even more rediculous, we stopped an m3.medium instance four weeks ago and starting this instance now fails.

We can solve this problem when switching from m3.medium to m4.large. Then everything works. But obviously m4.large is much more expensive than m3.medium.

Question: Is this problem with m3 instances a known problem?

What exactly is meant by "fails on boot"?

When we start the m3.medium instance it is never reachable via its IP address. In addition the AWS "Status Checks" in web console keep in state "Initializing". When we fetch an instance screenshot we see the last message "Booting from Hard Disk ..." but it never shows the Linux login prompt.

From my perspective it looks like a problem of the AMI with m3 instance type. In general m3 instance type is an old instance type which AWS want to have deprecated. But it is the only instance type with "medium" size. There is no m4.medium supported by Sophos UTM (regarding AWS Marketplace [2]).

[1] used AMI - aws-marketplace/sophos_utm_standalone_9.714-4.1-mp-9823791a-d242-49d1-8833-0a674195d6e2

[2] aws.amazon.com/.../prodview-gekyqjee5hefo



This thread was automatically locked due to age.
Parents
  • +1

    I believe you're running into a known issue where on UTM 9.713/9.714, some AWS instances can fail to boot. 

    This has been resolved in 9.715, and the fix was also pushed out in a pattern update (it's possible your instance didn't install the pattern update before updating to 9.713/9.714). 

    At this point, you have a few options: 

    1) Re-deploy your instance using 9.712, wait for pattern to update to latest, then upgrade to 9.713/9.714

    2) Re-deploy with 9.715 when it's available in AWS marketplace (it's currently in the process of being released to marketplace) 

    3) Continue to run as m4.large until you can upgrade to 9.715 

    There's also a way to manual recover these instances, but Sophos Support will have to apply it. If you contact them, you can reference NUTM-14015. 

  • 0 in reply to bobbylam

    We also have one 9.715 instance which has the same problem but I do not know whether the patterns are updated to latest.

    I will wait for official 9.715 AWS marketplace AMI and test it. As soon as the AMI is releases I will give you feedback.

  • 0 in reply to Martin Garbe

    Are your instances PV or HVM? https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/virtualization_types.html

    The UTM only support HVM types, so you may run into this issue as well if you're using PV instances. 

Reply Children
Unfiltered HTML