6.9 - RHSA-2009:1339

Problem Description:

An updated rgmanager package that fixes multiple security issues, various bugs, and adds enhancements is now available for Red Hat Enterprise Linux 5.

This update has been rated as having low security impact by the Red Hat Security Response Team.

2. Relevant releases/architectures:

RHEL Clustering (v. 5 server) - i386, ia64, ppc, x86_64

3. Description:

The rgmanager package contains the Red Hat Resource Group Manager, which provides high availability for critical server applications in the event of system downtime.

Multiple insecure temporary file use flaws were discovered in rgmanager and various resource scripts run by rgmanager. A local attacker could use these flaws to overwrite an arbitrary file writable by the rgmanager process (i.e. user root) with the output of rgmanager or a resource agent via a symbolic link attack. (CVE-2008-6552)

This update also fixes the following bugs:

* clulog now accepts '-' as the first character in messages.

* if expire_time is 0, max_restarts is no longer ignored.

* the SAP resource agents included in the rgmanager package shipped with Red Hat Enterprise Linux 5.3 were outdated. This update includes the most recent SAP resource agents and, consequently, improves SAP failover support.

* empty PID files no longer cause resource start failures.

* recovery policy of type 'restart' now works properly when using a resource based on ra-skelet.sh.

* samba.sh has been updated to kill the PID listed in the proper PID file.

* handling of the '-F' option has been improved to fix issues causing rgmanager to crash if no members of a restricted failover domain were online.

* the number of simultaneous status checks can now be limited to prevent load spikes.

* forking and cloning during status checks has been optimized to reduce load spikes.

* rg_test no longer hangs when run with large cluster configuration files.

* when rgmanager is used with a restricted failover domain it will no longer occasionally segfault when some nodes are offline during a failover event.

* virtual machine guests no longer restart after a cluster.conf update.

* nfsclient.sh no longer leaves temporary files after running.

* extra checks from the Oracle agents have been removed.

* vm.sh now uses libvirt.

* users can now define an explicit service processing order when central_processing is enabled.

* virtual machine guests can no longer start on 2 nodes at the same time.

* in some cases a successfully migrated virtual machine guest could restart when the cluster.conf file was updated.

* incorrect reporting of a service being started when it was not started has been addressed.

As well, this update adds the following enhancements:

* a startup_wait option has been added to the MySQL resource agent.

* services can now be prioritized.

* rgmanager now checks to see if it has been killed by the OOM killer and if so, reboots the node.

Users of rgmanager are advised to upgrade to this updated package, which resolves these issues and adds these enhancements.

4. Solution:

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

This update is available via Red Hat Network. Details on how to use the Red Hat Network to apply this update are available at http://kbase.redhat.com/faq/docs/DOC-11259

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

250718 - fs.sh inefficient scripting leads to load peaks and disk saturation 412911 - Convert all XM management calls to either lib virt or virsh 449394 - Recovery policy of type restart doesn't work with a service using a resource based on ra-skelet.sh 468691 - Virtual Services guest can start on 2 nodes at same time 470917 - The oracledb.sh script checks in strange intervals(10s, 5m, 4.5m) 471066 - rgmanager oracledb.sh resource agent does not properly check for all db startup failures. 471226 - oracledb.sh script kills ALL oracle instances when failing over 471431 - second ocf_log message doesn't make it to /var/log/messages 474444 - Zero-length pid files cause resource start failures 475826 - Update support for SAP resource agents (rgmanager) 481058 - MySQL Service Startup Timeout after Crash 482858 - Cluster Event Script needs Updates to include Group Exclusive 483093 - rgmanager: samba.sh tries to kill the wrong pid file 486349 - nfsclient.sh leaves temporary files /tmp/nfsclient-status-cache-$$ 486717 - clusvcadm -e <service> -F handling bugs 488714 - Enabling (according to failover domain rules) a frozen service results in a unusable failed+frozen service 489785 - /usr/share/cluster/apache.sh does not handle a valid /etc/httpd/conf/httpd.conf configuration correctly 490449 - domU's restart after cluster.conf update 490455 - rg_test hangs when running against cluster 492828 - RFE: priorities for services/virtual machines 494977 - segfault in check_rdomain_crash() during failover 505340 - VM migration and subsequent cluster.conf update can cause the VM restart 514044 - vm.sh does will fail resource if "no state" is detected 519436 - CVE-2008-6552 cman, gfs2-utils, rgmanager: multiple insecure temporary file use issues