5 - RHSA-2016:1853

Problem Description:

An update for atomic-openshift and heapster is now available for Red Hat OpenShift Enterprise 3.2.

Red Hat Product Security has rated this update as having a security impact of Important. A Common Vulnerability Scoring System (CVSS) base score, which gives a detailed severity rating, is available for each vulnerability from the CVE link(s) in the References section.

2. Relevant releases/architectures:

Red Hat OpenShift Enterprise 3.2 - noarch, x86_64

3. Description:

OpenShift Enterprise by Red Hat is the company's cloud computing Platform- as-a-Service (PaaS) solution designed for on-premise or private cloud deployments.

Security Fix(es):

* When processing an archive file that contains an archive entry with type 1 (hardlink) but also having a non-zero data size a file overwrite can occur. This would allow an attacker that can pass data to an application that uses libarchive to unpack it to overwrite arbitrary files with arbitrary data. (CVE-2016-5418)

Red Hat would like to thank Insomnia Security for reporting this issue.

This update also fixes the following bugs:

* Previously, pods that had a resource request of 0 and specified limits were classified as BestEffort when they should have been classified as Burstable. This bug fix ensures that those pods are correctly classified as Burstable.(BZ#1357475)

* Future versions of docker will require containerized installations of OpenShift Container Platform to mount /var/lib/origin with the `rslave` flag. New installations of OpenShift Container Platform 3.2 have this value set. However, upgrades from 3.1 did not properly set this value. This bug fix ensures that this flag is now set during upgrades, ensuring that OpenShift Container Platform works properly under future versions of docker. (BZ#1358197)

* The PersistentVolumeLabel admission plug-in is now enabled by default. This plug-in labels AWS and GCE volumes with their zone so the scheduler can limit the nodes for a pod to only those in the same zone as the persistent volumes being used by the pod. (BZ#1365600)

* Previously, heapster incorrectly generated error messages indicating that it "Failed to find node". This bug fix corrects that error and ensures that erroneous warnings are generated.(BZ#1366367)

* The deployment controllers' resync interval can now be configured. The previously hard-coded 2-minute default is the likely cause of performance regressions when thousands of deploymentconfigs are present in the system. Increase the resync interval by setting deploymentControllerResyncMinute in /etc/origin/master/master-config.yaml.(BZ#1366381)

* Previously, AWS-related environment variables were removed from /etc/sysconfig/atomic-openshift-master files during an upgrade if these values were not included in the advanced installer's inventory file. This bug fix ensures that these variables are now preserved during upgrades. (BZ#1370641)

* Previously, updates to the containerized atomic-openshift-node service were not properly reloaded during upgrades. This bug fix corrects this error and ensures that the service is reloaded during upgrades. (BZ#1371708)

* Previously the installer did not properly configure an environment for flannel when openshift_use_flannel was set to `true`. This bug fix corrects those errors and the installer will now correctly deploy environments using flannel. (BZ#1372026)

4. Solution:

For details on how to apply this update, which includes the changes described in this advisory, refer to:

https://access.redhat.com/articles/11258

The following images are included in this errata : openshift3/openvswitch:v3.2.1.15 openshift3/ose-pod:v3.2.1.15 openshift3/ose:v3.2.1.15 openshift3/ose-docker-registry:v3.2.1.15 openshift3/ose-keepalived-ipfailover:v3.2.1.15 openshift3/ose-recycler:v3.2.1.15 openshift3/ose-f5-router:v3.2.1.15 openshift3/ose-deployer:v3.2.1.15 openshift3/node:v3.2.1.15 openshift3/ose-sti-builder:v3.2.1.15 openshift3/ose-docker-builder:v3.2.1.15 openshift3/ose-haproxy-router:v3.2.1.15 openshift3/metrics-heapster:3.2.1-4

5. Bugs fixed (https://bugzilla.redhat.com/):

1357475 - Pod QoS Tier are different between OpenShift 3.2 and 3.3 1358197 - docker's per-mount propagation mode wasn't turn on after upgrade 1365600 - Volume affinity in OCP 3.2 1366367 - Heapster "Failed to find node" warning and verbose logging 1366381 - [ocp3.2.1] deployments and scale up/down are very, very slow 1370641 - Upgrade from 3.1 to 3.2 overwrites AWS variables in /etc/sysconfig/atomic-openshift-master-* 1371708 - atomic-openshift-node service wasn't reload on containerized env. 1372026 - openshift_use_flannel=true does not work properly