FAQ
Hello,

Trying to upload a stemcell to bosh micro running on AWS, I’m getting
the following error Error 100: The AMI ID 'ami-825ea7eb' does not
exist. Indeed, I can’t find the related ami within AWS.

I'm running a micro bosh instance running on the eu-west-1 zone, using
a AIM user (with ec2/allow/* policy) following Dr Nic's tutorial
uploading-public-stemcell.md after running creating-a-bosh-from-
scratch.md at https://github.com/drnic/bosh-getting-started. (I'm not
seeing the bosh micro commands though

I'm trying to debug this, but aws_registry is not outputting any debug
traces on the configured /var/vcap/deploy/bosh/aws_registry/shared/
logs/aws_registry.debug.log path. Any idea of how to diagnose and
solve this ?

Seems the 1.5.5 version of the aws-sdk could enable to output wire
traces to stdout
http://aws.amazon.com/releasenotes/Ruby/5234916478554933
# send wire traces to standard out
AWS.config(:http_wire_trace => true, :logger => nil)

Any guidance on steps to upgrade the aws-sdk on my bosh-micro to
leverage this ?

Full traces follow:

bosh upload stemcell bosh-stemcell-aws-0.5.1.tgz
[…]
Update stemcell
extracting stemcell archive (00:00:07)
verifying stemcell manifest (00:00:00)
checking if this stemcell already exists (00:00:00)
uploading stemcell bosh-stemcell/0.5.1 to the cloud: <?xml
version="1.0" encoding="UTF-8"?>
<Response><Errors><Error><Code>InvalidAMIID.NotFound</
Code><Message>The AMI ID 'ami-825ea7eb' does not exist</Message></
Error></Errors><RequestID>701e2b12-8fcc-4092-85bb-036a977a7f8b</
RequestID></Response> (00:09:26)
Error 4/5 00:09:33
Error 100: The AMI ID 'ami-825ea7eb' does not exist

~/stemcells$ bosh status
Updating director data... done

Target myfirstbosh (http://mydns.com:25555) Ver: 0.5.1
(33d8ae54)
UUID 22961c4a-6a22-4b4b-8215-cfc38e34f761
User admin
Deployment not set


Thanks in advance for your help,

Guillaume.

Search Discussions

  • Richard Lennox at Jun 30, 2012 at 5:12 am
    Hi,

    I had a similar issue. The error message is highly confusing - what
    it's actually complaining about is that it can't find aki-825ea7eb -
    which is a kernel image that is available in us-east-1a, but not in eu
    or ap zones. This is hard coded as a default in the AWS cpi
    (cloud.rb).

    I fixed this by untaring the public aws stemcell - and modifying the
    manifest file to

    ---
    name: bosh-stemcell
    bosh_protocol: "1"
    version: 0.5.1
    cloud_properties:
    kernel_id: "aki-fe1354ac"

    You'll need to find the right kernel image for the eu-west zone - that
    one above is ap-southeast. Then retar and gzip. Unload and away she
    goes. The kernel_id property overrides the DEFAULT_AKI (see cloud.rb
    for codebase).
    On Jun 30, 1:17 am, Guillaume Berche wrote:
    Hello,

    Trying to upload a stemcell to bosh micro running on AWS, I’m getting
    the following error Error 100: The AMI ID 'ami-825ea7eb' does not
    exist. Indeed, I can’t find the related ami within AWS.

    I'm running a micro bosh instance running on the eu-west-1 zone, using
    a AIM user (with ec2/allow/* policy) following Dr Nic's tutorial
    uploading-public-stemcell.md after running creating-a-bosh-from-
    scratch.md athttps://github.com/drnic/bosh-getting-started. (I'm not
    seeing the bosh micro commands though

    I'm trying to debug this, but aws_registry is not outputting any debug
    traces on the configured /var/vcap/deploy/bosh/aws_registry/shared/
    logs/aws_registry.debug.log path. Any idea of how to diagnose and
    solve this ?

    Seems the 1.5.5 version of the aws-sdk could enable to output wire
    traces to stdouthttp://aws.amazon.com/releasenotes/Ruby/5234916478554933
    # send wire traces to standard out
    AWS.config(:http_wire_trace => true, :logger => nil)

    Any guidance on steps to upgrade the aws-sdk on my bosh-micro to
    leverage this ?

    Full traces follow:

    bosh upload stemcell bosh-stemcell-aws-0.5.1.tgz
    […]
    Update stemcell
    extracting stemcell archive (00:00:07)
    verifying stemcell manifest (00:00:00)
    checking if this stemcell already exists (00:00:00)
    uploading stemcell bosh-stemcell/0.5.1 to the cloud: <?xml
    version="1.0" encoding="UTF-8"?>
    <Response><Errors><Error><Code>InvalidAMIID.NotFound</
    Code><Message>The AMI ID 'ami-825ea7eb' does not exist</Message></
    Error></Errors><RequestID>701e2b12-8fcc-4092-85bb-036a977a7f8b</
    RequestID></Response> (00:09:26)
    Error                   4/5 00:09:33
    Error 100: The AMI ID 'ami-825ea7eb' does not exist

    ~/stemcells$ bosh status
    Updating director data... done

    Target         myfirstbosh (http://mydns.com:25555) Ver: 0.5.1
    (33d8ae54)
    UUID           22961c4a-6a22-4b4b-8215-cfc38e34f761
    User           admin
    Deployment     not set

    Thanks in advance for your help,

    Guillaume.
  • Guillaume Berche at Jul 3, 2012 at 5:23 pm
    Thanks a lot Richard for your help!.

    With your advice, it went one step further and actually created an AMI (as
    observed in both AWS console and ec2 CLI), in addition to the 2GB EBS
    snapshot. But I'm still getting the "The AMI ID 'ami-1f16126b' does not
    exist" error message from bosh cli. Any idea ? Anyone tried with an IAM
    user yet ? Otherwise, I'm considering testing on us-east-1 zone instead of
    eu-west-1.

    How did you diagnose this aki issue Richard, did you manage to get
    additional traces somewhere ? Is it worth trying to get the aws-sdk wire
    traces to diagnose this ?

    More details:

    I tried with two akis (1.02 and 1.03 amd 64 x86). referenced into

    http://docs.amazonwebservices.com/AWSEC2/latest/UserGuide/UserProvidedkernels.html
    (version
    <http://docs.amazonwebservices.com/AWSEC2/latest/UserGuide/UserProvidedkernels.html>1.03)
    and *http://cloud-images.ubuntu.com/releases/oneiric/release/published-ec2-release.txt.orig
    (prevous 1.02 version)
    *<http://docs.amazonwebservices.com/AWSEC2/latest/UserGuide/UserProvidedkernels.html>

    the default aki-825ea7eb in bosh-stemcell-aws-0.5.1.tgz corresponds to:

    *us-east-1 aki-825ea7eb amd64 kernel
    ec2-public-images/pv-grub-hd0_1.02-amd64.gz.manifest.xml*


    I tried with the followings:

    *eu-west-1 aki-62695816 amd64 kernel ec2-public-images-eu/pv-grub-hd0_1.02-amd64.gz.manifest.xml*

    *aki-71665e05 ec2-public-images-eu/pv-grub-hd0_1.03-x86_64.gz.manifest.xml*


    **

    Thanks,

    Guillaume.
    **<http://docs.amazonwebservices.com/AWSEC2/latest/UserGuide/UserProvidedkernels.html>

    On Saturday, June 30, 2012 7:12:10 AM UTC+2, Richard Lennox wrote:

    Hi,

    I had a similar issue. The error message is highly confusing - what
    it's actually complaining about is that it can't find aki-825ea7eb -
    which is a kernel image that is available in us-east-1a, but not in eu
    or ap zones. This is hard coded as a default in the AWS cpi
    (cloud.rb).

    I fixed this by untaring the public aws stemcell - and modifying the
    manifest file to

    ---
    name: bosh-stemcell
    bosh_protocol: "1"
    version: 0.5.1
    cloud_properties:
    kernel_id: "aki-fe1354ac"

    You'll need to find the right kernel image for the eu-west zone - that
    one above is ap-southeast. Then retar and gzip. Unload and away she
    goes. The kernel_id property overrides the DEFAULT_AKI (see cloud.rb
    for codebase).
    On Jun 30, 1:17 am, Guillaume Berche wrote:
    Hello,

    Trying to upload a stemcell to bosh micro running on AWS, I’m getting
    the following error Error 100: The AMI ID 'ami-825ea7eb' does not
    exist. Indeed, I can’t find the related ami within AWS.

    I'm running a micro bosh instance running on the eu-west-1 zone, using
    a AIM user (with ec2/allow/* policy) following Dr Nic's tutorial
    uploading-public-stemcell.md after running creating-a-bosh-from-
    scratch.md athttps://github.com/drnic/bosh-getting-started. (I'm not
    seeing the bosh micro commands though

    I'm trying to debug this, but aws_registry is not outputting any debug
    traces on the configured /var/vcap/deploy/bosh/aws_registry/shared/
    logs/aws_registry.debug.log path. Any idea of how to diagnose and
    solve this ?

    Seems the 1.5.5 version of the aws-sdk could enable to output wire
    traces to stdouthttp://aws.amazon.com/releasenotes/Ruby/5234916478554933
    # send wire traces to standard out
    AWS.config(:http_wire_trace => true, :logger => nil)

    Any guidance on steps to upgrade the aws-sdk on my bosh-micro to
    leverage this ?

    Full traces follow:

    bosh upload stemcell bosh-stemcell-aws-0.5.1.tgz
    […]
    Update stemcell
    extracting stemcell archive (00:00:07)
    verifying stemcell manifest (00:00:00)
    checking if this stemcell already exists (00:00:00)
    uploading stemcell bosh-stemcell/0.5.1 to the cloud: <?xml
    version="1.0" encoding="UTF-8"?>
    <Response><Errors><Error><Code>InvalidAMIID.NotFound</
    Code><Message>The AMI ID 'ami-825ea7eb' does not exist</Message></
    Error></Errors><RequestID>701e2b12-8fcc-4092-85bb-036a977a7f8b</
    RequestID></Response> (00:09:26)
    Error 4/5 00:09:33
    Error 100: The AMI ID 'ami-825ea7eb' does not exist

    ~/stemcells$ bosh status
    Updating director data... done

    Target myfirstbosh (http://mydns.com:25555) Ver: 0.5.1
    (33d8ae54)
    UUID 22961c4a-6a22-4b4b-8215-cfc38e34f761
    User admin
    Deployment not set

    Thanks in advance for your help,

    Guillaume.
  • Guillaume Berche at Jul 3, 2012 at 8:52 pm
    Actually, I had missed Martin's "aws race condition workaround" open branch
    at http://reviews.cloudfoundry.org/#/c/6507/ that seems to match my
    symptoms, I'll try and confirm this fix the pb.

    Guillaume.
  • Martin Englund at Jul 4, 2012 at 12:29 am
    Guys,

    unfortunately I've been down with a cold for a while (still am bed-ridden, but finally well enough to read mail), and haven't kept an eye on the mailing list or I could have given you some hints...

    As you've seen, there is a race condition (I think I've uncovered at least one more) when it has finished uploading the stemcell to the block device and until it is available as an AMI, so as you noted, I've added an ugly workaround but am trying to figure out a cleaner way to do it.

    Cheers,
    /Martin

    Sent from my iPad
    On Jul 3, 2012, at 13:52, Guillaume Berche wrote:

    Actually, I had missed Martin's "aws race condition workaround" open branch at http://reviews.cloudfoundry.org/#/c/6507/ that seems to match my symptoms, I'll try and confirm this fix the pb.

    Guillaume.
  • Guillaume Berche at Jul 4, 2012 at 1:16 pm
    Thanks Martin for your email while sick in bed, I hope you recover well
    from your cold. Your fix indeed solved my symptoms, I can now proceed with
    deploying sample releases :-) thanks!

    Guillaume.
    On Wednesday, July 4, 2012 2:29:52 AM UTC+2, Martin Englund wrote:

    Guys,

    unfortunately I've been down with a cold for a while (still am bed-ridden,
    but finally well enough to read mail), and haven't kept an eye on the
    mailing list or I could have given you some hints...

    As you've seen, there is a race condition (I think I've uncovered at least
    one more) when it has finished uploading the stemcell to the block device
    and until it is available as an AMI, so as you noted, I've added an ugly
    workaround but am trying to figure out a cleaner way to do it.

    Cheers,
    /Martin

    Sent from my iPad
  • Hugues Malphettes at Jul 5, 2012 at 10:05 am
    I faced exactly the exact same issues than Guillaume.
    I confirm the fixes did bring me to the same point:
    - Specify the aki for my region in the stemcell's manifest
    - Apply Martin's patch to retry until the AMI is available
    The upload of the stemcell completed.
    Thanks again!
    Cheers,
    Hugues
    On Jul 4, 9:16 pm, Guillaume Berche wrote:
    Thanks Martin for your email while sick in bed, I hope you recover well
    from your cold. Your fix indeed solved my symptoms, I can now proceed with
    deploying sample releases :-)  thanks!

    Guillaume.






    On Wednesday, July 4, 2012 2:29:52 AM UTC+2, Martin Englund wrote:

    Guys,
    unfortunately I've been down with a cold for a while (still am bed-ridden,
    but finally well enough to read mail), and haven't kept an eye on the
    mailing list or I could have given you some hints...
    As you've seen, there is a race condition (I think I've uncovered at least
    one more) when it has finished uploading the stemcell to the block device
    and until it is available as an AMI, so as you noted, I've added an ugly
    workaround but am trying to figure out a cleaner way to do it.
    Cheers,
    /Martin
    Sent from my iPad
  • Bedzinski Waldemar at Jul 5, 2012 at 10:59 am
    2012/07/05//12:59//<w.r.bedzinski@gmail.com>
    thanks!
    On 5 July 2012 12:05, Hugues Malphettes wrote:

    I faced exactly the exact same issues than Guillaume.
    I confirm the fixes did bring me to the same point:
    - Specify the aki for my region in the stemcell's manifest
    - Apply Martin's patch to retry until the AMI is available
    The upload of the stemcell completed.
    Thanks again!
    Cheers,
    Hugues
    On Jul 4, 9:16 pm, Guillaume Berche wrote:
    Thanks Martin for your email while sick in bed, I hope you recover well
    from your cold. Your fix indeed solved my symptoms, I can now proceed with
    deploying sample releases :-) thanks!

    Guillaume.






    On Wednesday, July 4, 2012 2:29:52 AM UTC+2, Martin Englund wrote:

    Guys,
    unfortunately I've been down with a cold for a while (still am
    bed-ridden,
    but finally well enough to read mail), and haven't kept an eye on the
    mailing list or I could have given you some hints...
    As you've seen, there is a race condition (I think I've uncovered at
    least
    one more) when it has finished uploading the stemcell to the block
    device
    and until it is available as an AMI, so as you noted, I've added an
    ugly
    workaround but am trying to figure out a cleaner way to do it.
    Cheers,
    /Martin
    Sent from my iPad

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupbosh-users @
postedJun 29, '12 at 3:17p
activeJul 5, '12 at 10:59a
posts8
users5

People

Translate

site design / logo © 2022 Grokbase