5 - RHSA-2015:2290

Problem Description:

An updated pcs package that fixes one security issue, several bugs, and add various enhancements is now available for Red Hat Enterprise Linux 7.

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

2. Relevant releases/architectures:

Red Hat Enterprise Linux Server High Availability (v. 7) - s390x, x86_64 Red Hat Enterprise Linux Server Resilient Storage (v. 7) - s390x, x86_64

3. Description:

The pcs package provides a configuration tool for Corosync and Pacemaker. It permits users to easily view, modify and create Pacemaker based clusters. The pcs package includes Rack, which provides a minimal interface between webservers that support Ruby and Ruby frameworks.

A flaw was found in a way Rack processed parameters of incoming requests. An attacker could use this flaw to send a crafted request that would cause an application using Rack to crash. (CVE-2015-3225)

Red Hat would like to thank Ruby upstream developers for reporting this. Upstream acknowledges Tomek Rabczak from the NCC Group as the original reporter.

The pcs package has been upgraded to upstream version 0.9.143, which provides a number of bug fixes and enhancements over the previous version. (BZ#1198265)

The following enhancements are described in more detail in the Red Hat Enterprise Linux 7.2 Release Notes, linked to from the References section:

* The pcs resource move and pcs resource ban commands now display a warning message to clarify the commands' behavior (BZ#1201452)

* New command to move a Pacemaker resource to its preferred node (BZ#1122818)

This update also fixes the following bugs:

* Before this update, a bug caused location, ordering, and colocation constraints related to a resource group to be removed when removing any resource from that group. This bug has been fixed, and the constraints are now preserved until the group has no resources left, and is removed. (BZ#1158537)

* Previously, when a user disabled a resource clone or multi-state resource, and then later enabled a primitive resource within it, the clone or multi-state resource remained disabled. With this update, enabling a resource within a disabled clone or multi-state resource enables it. (BZ#1218979)

* When the web UI displayed a list of resource attributes, a bug caused the list to be truncated at the first "=" character. This update fixes the bug and now the web UI displays lists of resource attributes correctly. (BZ#1243579)

* The documentation for the "pcs stonith confirm" command was not clear. This could lead to incorrect usage of the command, which could in turn cause data corruption. With this update, the documentation has been improved and the "pcs stonith confirm" command is now more clearly explained. (BZ#1245264)

* Previously, if there were any unauthenticated nodes, creating a new cluster, adding a node to an existing cluster, or adding a cluster to the web UI failed with the message "Node is not authenticated". With this update, when the web UI detects a problem with authentication, the web UI displays a dialog to authenticate nodes as necessary. (BZ#1158569)

* Previously, the web UI displayed only primitive resources. Thus there was no way to set attributes, constraints and other properties separately for a parent resource and a child resource. This has now been fixed, and resources are displayed in a tree structure, meaning all resource elements can be viewed and edited independently. (BZ#1189857)

In addition, this update adds the following enhancements:

* A dashboard has been added which shows the status of clusters in the web UI. Previously, it was not possible to view all important information about clusters in one place. Now, a dashboard showing the status of clusters has been added to the main page of the web UI. (BZ#1158566)

* With this update, the pcsd daemon automatically synchronizes pcsd configuration across a cluster. This enables the web UI to be run from any node, allowing management even if any particular node is down. (BZ#1158577)

* The web UI can now be used to set permissions for users and groups on a cluster. This allows users and groups to have their access restricted to certain operations on certain clusters. (BZ#1158571)

All pcs users are advised to upgrade to this updated package, which corrects these issues and add these enhancements.

4. Solution:

Before applying this update, make sure all previously released errata relevant to your system have been applied.

For details on how to apply this update, refer to:

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

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

1121791 - Provide documentation of batch-limit and other pacemaker properties in man page or pcs help 1134426 - pcs needs a better parser for corosync.conf 1148863 - Pcsd backward/forward compatibility issues 1158491 - 'pcs cluster status' is documented to be an alias to 'pcs status cluster' but has different output 1158537 - Removing a resource from a group also removes constraints mentioning that group 1158571 - user and group support in gui - permissions to clusters managed by pcsd 1163671 - [RFE] Default corosync configuration should log to file 1163682 - nodes authentication stops if failed on one node 1165803 - pcs CLI should recognize and act upon "fail due to lack of authentication" state if/as suitable (e.g. for "pcs config restore") 1166160 - 'pcs acl role create' does not check syntax properly 1170205 - pcs cluster auth --force doesn't overwrite /var/lib/pcsd/tokens if its content is corrupt 1175400 - pcs resource op add creates duplicate op entires 1176687 - Pacemaker resource defaults should show up in 'pcs config' output 1182119 - A cloned resource banned on one of the nodes is shown as Inactive in GUI 1182793 - When attempting to add a duplicate fence level we get a non-useful error message 1182986 - Unable to find out value for require-all parameter for ordering constraint with clones 1183752 - Unable to delete VirtualDomain resource remote-node when it has configured some constraints 1185096 - debug-promote implementation 1186692 - cluster node removal should verify possible loss of quorum 1187320 - Uncloning a non-cloned resource produces invalid CIB 1187571 - ungrouping a resource from a cloned group produces invalid CIB when other resources exist in that group 1188571 - The --wait functionality implementation needs an overhaul 1189857 - need a tree view for clones/MS/groups in the resource panel [GUI] 1196412 - pcs cluster start should go to pcsd if user is not root 1197758 - pcs does not inform about incorrect command usage (pcs constraint order set) 1198222 - pcsd: GUI fails if orphaned resource is present in a cluster 1198265 - PCS Rebase bug for 7.2 1198274 - pcsd: don't automatically use --force everytime a resource is being removed 1198640 - [WebUI] spaces not allowed in resource agent options fields 1199073 - creating a resource name colliding with an existing group/clone/master ID needs better error message 1202457 - Referencing a non-existent ACL role should error out more gracefully 1204880 - pcs: stonith level value checking 1205653 - pcsd gui is not able to remove constraints and standby/unstandby nodes of remote cluster 1206214 - Formatting of longdesc metadata of resource agent is destroyed when using "pcs resource describe" 1206219 - pcs stonith describe only lists parameters of fence agent, but not description 1207805 - Need a way for pcs to clear out auth tokens 1212904 - better integration with standalone (unbundled) clufter package for cluster configuration conversion 1213429 - Cluster request fails on first node if this is not authorized 1215198 - pcsd: GUI ignores timeout value in fence_xvm agent form 1219574 - [gui] resource optional arguments: quoted strings missing 1231987 - pcs ought to require psmisc package (hidden dependency for killall execution) 1232292 - CVE-2015-3225 rubygem-rack: Potential Denial of Service Vulnerability in Rack normalize_params() 1235022 - Nagios metadata is missing 1247818 - pcs depends on initscripts 1250720 - traceback when running 'pcs resource enable clvmd --wait' 1253491 - pcs status pcsd shows "Unable to authenticate" on serial console 1257369 - pcs should print the output of crm_resource from pcs resource cleanup commands 1258619 - Ruby traceback on pcsd startup - /webrick.rb:48:in `shutdown': undefined method `shutdown' 1265425 - pcs is not parsing the output of crm_node properly 1268801 - A change in "crm_resource --set-parameter is-managed" introduces regression for Clone and M/S resources