VMware Cloud Community
dude64
Enthusiast
Enthusiast
Jump to solution

Failed to create OpenStack deployment. Not enough storage

Target management and compute clusters have 2 TB free, yet I can't get VIO 2.5 to deploy. Error message in wizard is "Failed to create OpenStack deployment. Not enough storage for (VIO-DB-0). At least 552GB storage is required to place all nodes...". In my case, both clusters already have VM's deployed and release notes indicate that resources available may not be reported correctly.  Any thoughts? Thanks, dude64

Tags (2)
Reply
0 Kudos
1 Solution

Accepted Solutions
dude64
Enthusiast
Enthusiast
Jump to solution

I do have 8 CPUs per ESXi host. I found that it was getting hung on the shared storage available within the 3 node cluster. Installation appears to use local storage only for DB nodes, and prefer to shared storage exists for all other nodes in my configuration. My shared storage originally was below the required 552 GB that VIO deployment was looking for. Once I replaced the shared storage with 1TB of shared storage, and deleted the old shared storage devices, it finally proceeded to install. I grep'd the vio logs which were not really helpful in determining the cause which would have been nice to get a more specific error message, and better yet have better docs around the detailed system requirements, i.e. local versus shared storage. It'd also be nice to see a doc on the logging and ansible playbook info. I'll mark this as answered.

Thanks, dude64

View solution in original post

Reply
0 Kudos
2 Replies
JunW
Enthusiast
Enthusiast
Jump to solution

Please have a look and see if each ESXi host has at least 8 logical processors and it has been added into your management/api access network portgroups. We have a fix recently where we should report error about missing logical processors or network connectivity but we reported no enough storage.

In case your ESXi host does not have 8 logical processors, you can edit /opt/vmware/vio/etc/omjs.properties file from vio manager VM, and change the attribute value of "oms.vmsize.cpu.controller" from 8 to 4, for example, and then run 'sudo service oms restart'.

Thanks,

Jun

Reply
0 Kudos
dude64
Enthusiast
Enthusiast
Jump to solution

I do have 8 CPUs per ESXi host. I found that it was getting hung on the shared storage available within the 3 node cluster. Installation appears to use local storage only for DB nodes, and prefer to shared storage exists for all other nodes in my configuration. My shared storage originally was below the required 552 GB that VIO deployment was looking for. Once I replaced the shared storage with 1TB of shared storage, and deleted the old shared storage devices, it finally proceeded to install. I grep'd the vio logs which were not really helpful in determining the cause which would have been nice to get a more specific error message, and better yet have better docs around the detailed system requirements, i.e. local versus shared storage. It'd also be nice to see a doc on the logging and ansible playbook info. I'll mark this as answered.

Thanks, dude64

Reply
0 Kudos