How to fix a failing VM if the actual assigned IP and the registered IP in BOSH database for the VM is not identical
Article ID: 293422
Updated On:
Products
Operations Manager
Issue/Introduction
Some of BOSH managed VMs (such as diego_cell VMs in the case of TAS) may fail because the actual assigned IP and registered IP in BOSH database for each VM is not identical while they should be identical.
For examples, suppose that you have two VMs called VM1 and VM2. In this example, suppose that 10.0.16.11 is allocated to VM1 and 10.0.16.12 is allocated to VM2 as actual IPs. And, suppose that VM1 is affected with this issue.
When this issue happens based on this example, you may see those assigned IPs are as below via bosh vms command or bosh instances command.
This article instructs how to fix those unexpected IP allocations.
For examples, suppose that you have two VMs called VM1 and VM2. In this example, suppose that 10.0.16.11 is allocated to VM1 and 10.0.16.12 is allocated to VM2 as actual IPs. And, suppose that VM1 is affected with this issue.
When this issue happens based on this example, you may see those assigned IPs are as below via bosh vms command or bosh instances command.
VM1 10.0.16.12 VM2 10.0.16.12In this situation, VM1's status via bosh vms command or bosh instances command is failing. If executing Apply Changes from Ops Manager in this situation, it fails because the IP is conflicted between VM1 and VM2 according to registered IPs in BOSH database.
This article instructs how to fix those unexpected IP allocations.
Resolution
This issue happens when some of BOSH database columns storing IP information is corrupted. The basic idea is to recreate the affected VMs and remove those corrupted columns as a result.
If you see the issue after scaling out the number of VMs in the specific VM group (such as, diego_cell in TAS case) and finishing Apply Changes from Ops Manager, then please try to revert back the number of VMs and execute Apply Changes from Ops Manager. This operation should remove the affected VMs as a result.
Otherwise, please try to recreate those affected VMs with using bosh recreate command WITHOUT --no-converge.
If you can't resolve the issue with applying the above operations (in many cases, you fail to finish bosh recreate command itself with some errors), please contact VMware Tanzu support.
If you see the issue after scaling out the number of VMs in the specific VM group (such as, diego_cell in TAS case) and finishing Apply Changes from Ops Manager, then please try to revert back the number of VMs and execute Apply Changes from Ops Manager. This operation should remove the affected VMs as a result.
Otherwise, please try to recreate those affected VMs with using bosh recreate command WITHOUT --no-converge.
If you can't resolve the issue with applying the above operations (in many cases, you fail to finish bosh recreate command itself with some errors), please contact VMware Tanzu support.
Feedback
Yes
No
Powered by
