search

vmware-archive / cfops

This is simply an automation that is based on the supported way to back up Pivotal Cloud Foundry
http://www.cfops.io
Apache License 2.0
35 stars 24 forks source link

Stemcells are not restored #77

Closed ebornier-pivotal closed 7 years ago

ebornier-pivotal commented 8 years ago

Hi,

Stemcells imported into ops manager are not restored after a backup / restore process from scratch (with a fresh new ops manager).

Erwan Bornier

xchapter7x commented 8 years ago

cfops backup/restore for ops manager simply automates the import/export functionality available in the ops manager web UI. This is simply how ops manager functions.

I believe the reason for this is, restore is intended just to restore the ops manager config. It is likely that the restore is being done from a archive generated on a different version of ops manager. As such it would require a different stemcell potentially.

We do have a spike to explore backing up / restoring the director vm and with it, the bosh blobstore & db. This would restore the state of the available stemcells as that is where they are stored. This should address the stemcell issue you are raising.

Again, this is simply how ops manager import/export works. Since we merely automate that via a cli utility, i dont believe there is an issue to be resolved here.

I will keep you posted on the Director VM backup/restore functionality mentioned above.

ebornier-pivotal commented 8 years ago

Hi John,

Thanks you for your feedback.

Yes, my request was more an enhancement than an issue. It was expressed by my customer (Renault) during a ops dojo, it was perceived like something missing.

Happy to know that this could be address by a more global user story.

My request was in the context of disaster recovery with a ops manager recreation from scratch -- same ops manager version.

Regards,

Erwan

On Tue, Apr 26, 2016 at 9:24 AM, John Calabrese notifications@github.com wrote:

cfops backup/restore for ops manager simply automates the import/export functionality available in the ops manager web UI. This is simply how ops manager functions.

I believe the reason for this is, restore is intended just to restore the ops manager config. It is likely that the restore is being done from a archive generated on a different version of ops manager. As such it would require a different stemcell potentially.

We do have a spike to explore backing up / restoring the director vm and with it, the bosh blobstore & db. This would restore the state of the available stemcells as that is where they are stored. This should address the stemcell issue you are raising.

Again, this is simply how ops manager import/export works. Since we merely automate that via a cli utility, there isnt any issue to be resolved here.

I will keep you posted on the Director VM backup/restore functionality mentioned above.

— You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub https://github.com/pivotalservices/cfops/issues/77#issuecomment-214801118

Erwan Bornier | Cloud Platform Architect, Southern EMEA | Pivotal +33 6 62 23 31 90 | ebornier@pivotal.io

matthewcosgrove commented 8 years ago

Just going through this with a customer, and wanted to propose that there is potential approach which is scriptable and could therefore be incorporated into cfops. Essentially, as I understand it, we need to trick (is 'educate' a better word?) OpsManager into believing this is a fresh install. This can be achieved by deleting the following file from the OpsManager VM prior to clicking Apply Changes

/var/tempest/workspaces/default/deployments/bosh-state.json