Saksham Srivastava created CLOUDSTACK-6427:
----------------------------------------------
Summary: nic_count not decremented in case of failed VM start
Key: CLOUDSTACK-6427
URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6427
Project: CloudStack
Issue Type: Bug
Security Level: Public (Anyone can view this level - this is the default.)
Components: Management Server
Affects Versions: 4.4.0
Environment: 4.4.0, Xenserver
Reporter: Saksham Srivastava
Assignee: Saksham Srivastava
Fix For: 4.4.0
nics_count (op_networks) table is incremented on every user VM start and
decremented on every user VM stop. Essentially, nics_count associated with a
network represents if there are any running VMs using that network.
In case VM start fails we initiate a cleanup where in first we send the
stopCommand to the VM and upon success we release the network resources (in
DB). nics_count is decremented during this process. However, if the stopCommand
fails for some reason the resources are not released. Hence the nics_count is
not decremented.
At this point the nics_count gets out of sync, that is, it is greater than the
number of running VMs. This can result in the network never being GC'ed.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
