Thursday, January 29, 2015

Exadata Ghost Vulnerability

There is a new Oracle Linux vulnerability called Ghost documented in cve-2015-0235 this affects Oracle Exadata.

As of yet the Oracle Security and Patch patch page should be updated soon.

http://www.oracle.com/technetwork/topics/security/alerts-086861.html


The solution for Exadata is well documented in the support note below.

glibc vulnerability (CVE-2015-0235) patch availability for Oracle Exadata Database Machine (Doc ID 1965525.1)


Download and stage the files needed
For Exadata image versions 12.1.1.1.1 or earlier, obtain updated packages using the following package versions, or later package versions, if available:
glibc-2.5-123.0.1.el5_11.1.i686.rpm
glibc-2.5-123.0.1.el5_11.1.x86_64.rpm
glibc-common-2.5-123.0.1.el5_11.1.x86_64.rpm
glibc-devel-2.5-123.0.1.el5_11.1.i386.rpm
glibc-devel-2.5-123.0.1.el5_11.1.x86_64.rpm
glibc-headers-2.5-123.0.1.el5_11.1.x86_64.rpm
glibc-utils-2.5-123.0.1.el5_11.1.x86_64.rpm
nscd-2.5-123.0.1.el5_11.1.x86_64.rpm
For Exadata image version 12.1.2.1.0, obtain updated packages using the following package versions, or later package versions, if available:
glibc-2.12-1.149.el6_6.5.i686.rpm
glibc-2.12-1.149.el6_6.5.x86_64.rpm
glibc-common-2.12-1.149.el6_6.5.x86_64.rpm
glibc-devel-2.12-1.149.el6_6.5.i686.rpm
glibc-devel-2.12-1.149.el6_6.5.x86_64.rpm
glibc-headers-2.12-1.149.el6_6.5.x86_64.rpm
nscd-2.12-1.149.el6_6.5.x86_64.rpm
Oracle Exadata Database Servers running Linux
To install these packages on database servers, follow the steps below (applicable to all Exadata image versions). These may be done in parallel on all database servers or in a rolling manner. It is intended that the package installation is followed by a system reboot in a relatively short time (i.e. minutes, not days). Since the system will be rebooted, you may choose to stop the database and cluster processes on the node being updated in advance or allow the reboot process to stop them for you. 
1.  Capture the currently installed rpm versions (including package architectures) using the following command and save the output in a file in case a rollback is needed later.
1.                    rpm -qa --queryformat="%{name}-%{version}-%{release}.%{arch}\n" | egrep 'glibc|nscd'
2.  Stage the files on each database server in /tmp/glibc-update as root
1.                    mkdir /tmp/glibc-update
2.                    Place all the rpms listed above (for your appropriate release) in the directory /tmp/glibc-update
3.  If using Exadata Database Server image version 11.2.3.3.0 or later, run this command
1.                    rpm -e exadata-sun-computenode-exact
4.  For all releases: install the updated rpms using this command
1.                    rpm -Fvh /tmp/glibc-update/*rpm
5.  If the installation is successful (no errors), reboot the system using
1.                    shutdown -r -y now
6.  After the reboot, ensure the system is up and running and the cluster processes have restarted. Remove the staged files, if desired
1.                    rm -rf /tmp/glibc-update
If a rollback is required, it should be done with Oracle Support guidance via an SR. The information gathered in step 1 above should be provided to the SR. 




Note that it is not necessary to relink any binaries after this update. 
Oracle Exadata Database Servers running Solaris
Solaris systems do not include glibc.
Oracle Exadata Storage Cells
For storage cells, obtain the same files listed above for database servers and follow these steps for installation on the storage cells. While storage cells are not normally permitted to have OS updates applied, this procedure is allowed as an exception to address this vulnerability only. 
To install these updates, the storage cell will need to be rebooted. This can be done in a rolling manner in order to minimize availability impact to the system. Before attempting the installation procedures below, it is recommended to review Note 1188080.1 for procedures to gracefully take a storage cell offline before rebooting it and then bringing it back online after the reboot.  
If desired, all cells can be done in parallel as long as the cluster is shutdown before rebooting the cells. 
Note: Do not remove the exadata-sun-cellnode-exact package on storage cells.
To install these packages on storage cells, follow the steps below (applicable to all Exadata image versions).
1.  Capture the currently installed rpm versions (including package architectures) using the following command and save the output in a file in case a rollback is needed later.
1.                    rpm -qa --queryformat="%{name}-%{version}-%{release}.%{arch}\n" | egrep 'glibc|nscd'
2.  Stage the files on each storage cell in /tmp/glibc-update as root
1.                    mkdir /tmp/glibc-update
2.                    Place all the rpms listed above (for your appropriate release) in the directory /tmp/glibc-update
3.  Install the updated rpms using this command
1.                    rpm -Fvh --nodeps /tmp/glibc-update/*rpm
4.  If the installation is successful (no errors), reboot the system using
1.                    shutdown -r -y now
5.  After the reboot, ensure the system is up and running. Remove the staged files, if desired
1.                    rm -rf /tmp/glibc-update
6.  Follow the steps from Note 1188080.1 to ensure the cell is fully online again before proceeding to the next storage cell. 
If a rollback is required, it should be done with Oracle Support guidance via an SR. The information gathered in step 1 above should be provided to the SR. 

No comments:

Post a Comment