yahoo-eng-team team mailing list archive

[Matt Riedemann <mriedem@xxxxxxxxxx>]

This isn't a bug, you're running in a non-default environment for the
project which requires you to make changes to the the configuration
files for it to work.  That doesn't mean we should change the default
values in the conf file to accommodate a scale test.  If there is
something else being missed here or should be looked at deeper by the
project team, it should really go to oslo as this is an oslo change, not
a nova change.

** Changed in: nova
       Status: New => Invalid

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1225655

Title:
  rpc time out when create 100/200 VMs in one request

Status in OpenStack Compute (Nova):
  Invalid

Bug description:
  When create 100/200 VMs in one request, many VMs create failed caused
  by rpc time out.

  The root cause is that rpc_conn_pool_size and pc_response_timeout is
  not big enough to work well in large scale clusters.

      cfg.IntOpt('rpc_conn_pool_size',
                 default=30,
                 help='Size of RPC connection pool'),
      cfg.IntOpt('rpc_response_timeout',
                 default=60,
                 help='Seconds to wait for a response from call or multicall'),

  Based on some test, suggest change the default value as following:

  rpc_conn_pool_size ==> 100
  rpc_response_timeout ==> 1200

  This can make sure that user do not need to update nova.conf when they
  are using large scale cluster.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1225655/+subscriptions