What is "SSD backed PD Capacity"?

[Michael Dillon]

I have a charge on my account that doesn't seem to be reflected in the cost estimate calculators:

I have no running instances using SSD, but there's a flat fee listed as 'SSD backed PD Capacity'. I say 'flat' because it's been the same fee for the last two months - it doesn't fluctuate with usage.

What is it? How can I reduce it?

Thanks!

Mike

[Kamran (Google Cloud Support)]

Hello Michael,

Please use billing support requests form to get in touch with Google Cloud billing team and to get answers to your billing questions.

Sincerely,

[Scott Van Woudenberg]

Hi Michael,

You indicated that you have no running instances using PD SSD, which makes me think that you may have a stopped instance in your project? If so, that would explain the charge. Stopped VMs no longer generate the $/hr cost of a running VM, but PDs and PD SSDs associated with a stopped instance are still billed at the normal block storage rates. Refer to this docs page for details.

As far as options to reduce this cost, you can create a snapshot of the PD SSD volume and delete the stopped instance (snapshot storage is quite a bit cheaper than PD SSD) and associated PD SSD volume. If the PD SSD is your instance's boot disk, and you want to make it easier to recreate the instance, you can create a custom image (image storage, while not as cheap as snapshot storage, is still cheaper than PD SSD storage) from the PD SSD boot volume before deleting the volume itself.

Of course I'm just guessing. If I'm wrong, then by all means, please use the link that Kamran provided below to get our billing team involved to help you sort this out.

HTH,

-ScottVW

---

Scott Van Woudenberg

Product Manager

Google Compute Engine

--
© 2016 Google Inc. 1600 Amphitheatre Parkway, Mountain View, CA 94043
 
Email preferences: You received this email because you signed up for the Google Compute Engine Discussion Google Group (gce-discussion@googlegroups.com) to participate in discussions with other members of the Google Compute Engine community and the Google Compute Engine Team.
---
You received this message because you are subscribed to the Google Groups "gce-discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gce-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to gce-dis...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gce-discussion/a1411d98-bdde-4717-9bdd-06d09d12c39f%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Theodore Ts'o]

On Sun, Sep 11, 2016 at 11:53:00AM -0700, 'Scott Van Woudenberg' via gce-discussion wrote:
>
> You indicated that you have no *running* instances using PD SSD, which

> makes me think that you may have a stopped instance in your project? If so,
> that would explain the charge. Stopped VMs no longer generate the $/hr cost
> of a running VM, but PDs and PD SSDs associated with a stopped instance are
> still billed at the normal block storage rates

> <https://cloud.google.com/compute/pricing#disk>. Refer to this docs page
> <https://cloud.google.com/compute/docs/instances/stopping-or-deleting-an-instance#billing_for_stopped_instances>
> for details.

It's also possible that on one or more running instances, the disk's
"auto delete" flag was cleared[1] --- or depending on how the disk was
attached to the instance, was never set in the first place. In that
case, the SSD-backed persistent disk could have survived the death of
the instance(s). If you have the gcloud command-line SDK installed, a
useful command to list the disks currently associated with your GCE
project is "gcloud compute disks list"[2], which might result in a
listing like this:

NAME ZONE SIZE_GB TYPE STATUS
freebsd-10 us-central1-c 22 pd-standard READY
freebsd-11 us-central1-c 22 pd-ssd READY

So in the above listing you will see that there is 22GB of pd-standard
storage, and 22GB of pd-ssd storage. These disks can exist even if
they are not attached to either a running or stopped instance. If
they are not currently attached to an instance, you can either delete
the disk[3], if you are sure you do not need it, or you can attach the
disk[4] to a running instance if you want to take a look at it before
deciding what to do with it.

And of course, you can also do any of these operations using the web
interface via the Cloud Console[5] or via any of the Google API
clients[6]. Personally, I find the gcloud SDK[7] interface the most
convenient, since I can easily create shell scripts using the command
line API, but depending on your preferences and what you are trying to
do, other alternatives may be better for you.

Hope this helps,

- Ted

[1] https://cloud.google.com/sdk/gcloud/reference/compute/instances/set-disk-auto-delete
[2] https://cloud.google.com/sdk/gcloud/reference/compute/disks/list
[3] https://cloud.google.com/sdk/gcloud/reference/compute/disks/delete
[4] https://cloud.google.com/sdk/gcloud/reference/compute/instances/attach-disk
[5] https://console.cloud.google.com/compute/instances
[6] https://developers.google.com/api-client-library/
[7] https://cloud.google.com/sdk/

[Michael Dillon]

Thanks Kamran - 

I'd tried that first - but the response I received was honestly unhelpful.  And the suggestion I received (export your reports) was not really useful. 

The precise response from Scott (Stopped VMs no longer generate the $/hr cost of a running VM, but PDs and PD SSDs associated with a stopped instance are still billed at the normal block storage rates) was, I think, the root cause.

Mike

[Michael Dillon]

Thanks Scott!

Yes - I think my stopped instance (a 200GB SSD disk) was the cause.  It was costing $34/month.  That was a surprise!

Mike

On Sunday, September 11, 2016 at 12:53:04 PM UTC-6, Scott Van Woudenberg wrote:

Hi Michael,

You indicated that you have no running instances using PD SSD, which makes me think that you may have a stopped instance in your project? If so, that would explain the charge. Stopped VMs no longer generate the $/hr cost of a running VM, but PDs and PD SSDs associated with a stopped instance are still billed at the normal block storage rates. Refer to this docs page for details.

As far as options to reduce this cost, you can create a snapshot of the PD SSD volume and delete the stopped instance (snapshot storage is quite a bit cheaper than PD SSD) and associated PD SSD volume. If the PD SSD is your instance's boot disk, and you want to make it easier to recreate the instance, you can create a custom image (image storage, while not as cheap as snapshot storage, is still cheaper than PD SSD storage) from the PD SSD boot volume before deleting the volume itself.

Of course I'm just guessing. If I'm wrong, then by all means, please use the link that Kamran provided below to get our billing team involved to help you sort this out.

HTH,

-ScottVW

---

Scott Van Woudenberg

Product Manager

Google Compute Engine

On Sun, Sep 11, 2016 at 11:28 AM, 'Kamran (Google Cloud Support)' via gce-discussion <gce-dis...@googlegroups.com> wrote:

Hello Michael,

Please use billing support requests form to get in touch with Google Cloud billing team and to get answers to your billing questions.

Sincerely,

On Saturday, September 10, 2016 at 3:17:44 PM UTC-4, Michael Dillon wrote:

I have a charge on my account that doesn't seem to be reflected in the cost estimate calculators:

I have no running instances using SSD, but there's a flat fee listed as 'SSD backed PD Capacity'. I say 'flat' because it's been the same fee for the last two months - it doesn't fluctuate with usage.

What is it? How can I reduce it?

Thanks!

Mike

--
© 2016 Google Inc. 1600 Amphitheatre Parkway, Mountain View, CA 94043
 

Email preferences: You received this email because you signed up for the Google Compute Engine Discussion Google Group (gce-dis...@googlegroups.com) to participate in discussions with other members of the Google Compute Engine community and the Google Compute Engine Team.

---
You received this message because you are subscribed to the Google Groups "gce-discussion" group.

To unsubscribe from this group and stop receiving emails from it, send an email to gce-discussio...@googlegroups.com.

To post to this group, send email to gce-dis...@googlegroups.com.

[Paul Nash]

Hi Mike, glad to hear it got sorted out.

Unfortunately this is a common gotcha in public cloud (any cloud). The storage behind disks used by virtual machines costs money to maintain, regardless of whether the VM itself is running, and is billed until it is deleted. If the contents of the disk are not important to you once the VM has been stopped, we do offer the "delete on delete" option which will delete the disk when the VM is *deleted*. But not when it is merely stopped. It's like the difference between powering down your PC for the night, vs. chucking it in the dumpster. :)

Email preferences: You received this email because you signed up for the Google Compute Engine Discussion Google Group (gce-discussion@googlegroups.com) to participate in discussions with other members of the Google Compute Engine community and the Google Compute Engine Team.

---
You received this message because you are subscribed to the Google Groups "gce-discussion" group.

To unsubscribe from this group and stop receiving emails from it, send an email to gce-discussion+unsubscribe@googlegroups.com.
To post to this group, send email to gce-discussion@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gce-discussion/c0412086-1682-43f6-8571-a1fb26b93fd3%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--

Paul R. Nash |

Product Manager, Compute Engine |

paul...@google.com |

206-876-1620

[Hoa Le]

Hi Ted,

I tried to use [3] but it said 

ERROR: (gcloud.compute.disks.delete) Could not fetch resource:

And looks like I'm still getting charge everyday

[Hoa Le]

@here, anyone knows how to delete all the resources that are not found:

gcloud compute disks delete $(gcloud compute disks list | awk '{print $1}') --quiet

ERROR: (gcloud.compute.disks.delete) Could not fetch resource:

 - The resource 'projects/{myproject}/zones/us-central1-a/disks/NAME' was not found

 - The resource 'projects/{myproject}/zones/us-central1-a/disks/gke-cluster-1-0378a6c1-pvc-1b40276d-49c7-11e8-bc84-42010a800169' was not found

 - The resource 'projects/{myproject}/zones/us-central1-a/disks/gke-cluster-1-0378a6c1-pvc-38e3bcdd-49c7-11e8-bc84-42010a800169' was not found

 - The resource 'projects/{myproject}/zones/us-central1-a/disks/gke-cluster-1-0378a6c1-pvc-5a78fc4a-49c7-11e8-bc84-42010a800169' was not found

 - The resource 'projects/{myproject}/zones/us-central1-a/disks/gke-cluster-1-0378a6c1-pvc-b5162cd4-49d1-11e8-bc84-42010a800169' was not found

 - The resource 'projects/{myproject}/zones/us-central1-a/disks/gke-cluster-1-0378a6c1-pvc-b516b9bf-49d1-11e8-bc84-42010a800169' was not found

[Piweling Ulama]

Hi,

i have same issue,

I use gcp free tier before, then i have charge for Compute Engine SSD backed PD Capacity after free tier ends.

I really need your guidance to solve my problem.

thanks,

Aceng

[Adebisi Ibirogba]

As previously stated, you might want to check if you still have " SSD backed PD " running as this would be billed even though the VM has been deleted.

[Piweling Ulama]

I don't know that there are still gcp services still running.
Billing support has solved my issue.  

[Ahmad P - Cloud Platform Support]

SSD PD stands for Solid State Drive Persistent disk. In this documentation you can see the storage options. When you configured your project, you must likely have selected a VM type, RAM and storage that might be the origin of the SSD charges.

Your project is using storage PD that is over the capacity awarded by the VM setup. So the additional resources needed are added as the second storage charge. Storage charges are charges on a monthly basis so the extra storage was set up with anticipation during the project configuration.

It seems your issue has resolved by Billing team.