mount point /var/vcap/data was based on ephemeral disk(/dev/sdb2) about bosh-azure-cpi-release HOT 9 CLOSED

You should not stop the VMs. Please reference this issue #55

from bosh-azure-cpi-release.

Thank you for your response.
I have read this issue, and I understand this is BOSH design.

I want to know why BOSH use ephemeral disk in /var/vcap/data
What should I do when I want to restart VM? (e.g. change VM size.)

from bosh-azure-cpi-release.

You can restart VM but cannot stop and then start it. If a VM is reallocated in a different host, you can use 'bosh cloudcheck' to repair it.

Under my understanding, about why BOSH puts /var/vcap/data in ephemeral disk,

1. Ephemeral disk is a local disk and has higher speed. Although OS disk is also local disk, the space of OS disk may be not enough.
2. Administrator often uses BOSH to update the deployment and the data under /var/vcap/data are also changed frequently.

from bosh-azure-cpi-release.

I can understand.

Thank you for your help.

from bosh-azure-cpi-release.

@toshiomakiyy
FYI.
If you want to change the VM size, you need to update instance_type in your manifest and re-deploy it. Then BOSH will delete the old VM and create a new one with a different VM size.

from bosh-azure-cpi-release.

Hi there

Just getting started with Cloud Foundry and hit the same issue after triggering a manual restart of the BOSH Director VM itself via the Azure Portal.

Am I correct in understanding that by design the BOSH Director VM will not recover from such a restart? I can understand the use of the local disk for performance reasons but I would expect the Direcor to be able to recover from such restarts will happen either manually or when Azure patches the VMs.

Thanks

from bosh-azure-cpi-release.

You can easily deploy it with bosh-init again. Your data is kept in persistent data disk.

Best Regards,
Abel HU

? 2016?3?11??20:32?yorkie-m <[email protected]mailto:[email protected]> ???

Hi there

Just getting started with Cloud Foundry and hit the same issue after triggering a manual restart of the BOSH Director VM itself via the Azure Portal.

Am I correct in understanding that by design the BOSH Director VM will not recover from such a restart? I can understand the use of the local disk for performance reasons but I would expect the Direcor to be able to recover from such restarts will happen either manually or when Azure patches the VMs.

Thanks

Reply to this email directly or view it on GitHubhttps://github.com//issues/71#issuecomment-195349135.

from bosh-azure-cpi-release.

Many thanks Abel. I've come back to this issue and on trying bosh-init deploy I get the below error. Investigating now but any help appreciated

Thanks
Mark.

Started validating
Failed validating (00:00:00)
Command 'deploy' failed:
Parsing release set manifest '~/example_manifests/single-vm-cf.yml':
Validating release set manifest:
releases[0].url must be a valid URL (file:// or http(s)://)

from bosh-azure-cpi-release.

@yorkie-m Did you find the reason for the invalid manifest? If not, you can send ~/example_manifests/single-vm-cf.yml to me ([email protected]).

from bosh-azure-cpi-release.