(RHSA-2014:0578) Moderate: openstack-nova security, bug fix, and enhancement update

OpenStack Compute (nova) launches and schedules large networks of virtual
machines, creating a redundant and scalable cloud computing platform.
Compute provides the software, control panels, and APIs required to
orchestrate a cloud, including running virtual machine instances, managing
networks, and controlling access through users and projects.

It was found that overwriting the disk inside of an instance with a
malicious image, and then switching the instance to rescue mode, could
potentially allow an authenticated user to access arbitrary files on the
compute host depending on the file permissions and SELinux constraints of
those files. Only setups that used libvirt to spawn instances and which had
the use of cow images disabled (“use_cow_images = False” in nova
configuration) were affected. (CVE-2014-0134)

Red Hat would like to thank the OpenStack Project for reporting this issue.
Upstream acknowledges Stanislaw Pitucha from Hewlett Packard as the
original reporter.

Bug fixes:

• The Compute libguestfs did not perform tear down correctly in certain
  cases, such as when attempting to perform SSH key injection when no SSH
  keys were specified. As a result, libguestfs qemu instances would remain
  running, even after the Compute instance with which they were associated
  was destroyed. This has been fixed and now tear down of libguestfs occurs
  correctly. (BZ#1022627)

• Qpid’s driver reconnection delay is not configurable. This and the fact
  that the hard-coded delay was quite high became a blocker issue from an HA
  perspective. Making this value configurable is not possible for this
  version, so the hard-coded delay was changed. The new delay cap is 5
  seconds. (BZ#1060772)

• Nova was creating a temporary VMware snapshot during a nova snapshot, but
  not deleting it on completion. Thus VMware guests could end up with a large
  number of unnecessary snapshots. The temporary VMware snapshot is now
  deleted on completion of the nova snapshot operation. (BZ#1069429)

• The Nova vmware driver was not passing the iscsi server location to the
  instance, hence it was not possible to add a cinder iscsi disk to a VMware
  instance. The nova vmware driver now correctly passes the iscsi server
  location to the instance and it is possible to add a cinder iscsi disk to a
  VMware instance. (BZ#1069432)

• The nova VMware driver was assigning VNC ports in a way which could cause
  collisions between guests. There was a small chance that 2 instances could
  be created with the same VNC port. If both instances ran on the same ESX
  host, only one of their consoles would be available. However, that console
  would be presented when requested for both instances. This has been fixed
  so that the nova VMware driver now allocates VNC ports which are unique to
  a vcenter, and a user of one instance will not be presented with the
  console of another. (BZ#1077017)

Enhancement:

• When troubleshooting production systems it is desirable to be able to
  trace all data base queries, web REST calls, messaging service RPC calls,
  and libvirt API calls associated with invocation of a user command or
  background job. Hence, the ability has been added to obtain a debug dump of
  the system state for an OpenStack service based process. For example, upon
  sending SIGUSR1, a dump of all native threads, green threads, live
  configuration, and any other relevant info is triggered. (BZ#1071469)

All openstack-nova users are advised to upgrade to these updated packages,
which correct these issues and add this enhancement.