Oracle Linux 5.9 Release Notes

Updated January 2013

Introduction

This document covers the following topics regarding Oracle Linux 5.9.

For the latest updates please refer to the online version of these release notes available at:

http://oss.oracle.com/el5/docs/

NOTE: An upgrade of Oracle Linux from a beta release is not supported. Further, an in-place upgrade between major versions of Oracle Linux is not supported. Oracle does not recommend an upgrade from earlier major versions of Oracle Linux even though anaconda provides an option to do this upgrade. A fresh installation is strongly recommended rather than a system upgrade between major versions.

Changes from the Upstream Release

The following RPMs are modified from the upstream release. All changes are trademark and look/feel related unless otherwise noted below under the specific RPM.

The following packages have been removed:

The following documentation packages have been removed:

New Packages added in this release

The following packages were added upstream for this release (5.9) which were not in the previous update release (5.8)

Packages added by Oracle

The following packages have been added to the release:

Kernel and Driver Updates

Oracle Linux 5.9 ships with the following kernel packages

Note:Oracle Linux 5.9 includes both a 32 bit and a 64 bit Unbreakable Enterprise Kernel. Both Unbreakable Enterprise kernel and Red Hat compatible kernel are installed and the system boots with Unbreakable Enterprise kernel by default. If needed /etc/grub.conf can be modified to make the system boot with Red Hat compatible kernel by default.

Red Hat Compatible Kernel

This section covers significant changes in the Red Hat compatible kernel in this release.

Driver Updates from the upstream distribution provider

Storage

Network

Miscellaneous

Red Hat Compatible Kernel with Bug Fixes by Oracle

This is the Red Hat compatible kernel with critical bug fixes produced by Oracle.

Bug Fixes

Installation

The Red Hat Compatible Kernel with Bug Fixes by Oracle must be manually installed. To install this kernel, follow these steps:

  1. Mount Oracle Linux 5.9 DVD
  2. Copy the RPMs from mount_point/Server/oracle_updated to your system
  3. Install required kernel using the rpm command, e.g.
    # rpm -ihv kernel-2.6.18-308.0.0.0.1.el5.x86_64.rpm

Unbreakable Enterprise Kernel

Introduction

The Unbreakable Enterprise Kernel Release 2 is Oracle's second major release of its heavily tested and optimized operating system kernel for Oracle Linux 5 and Oracle Linux 6. It is based on the mainline Linux 3.0 version 3.0.36. It contains a large number of improvements and new features that have been incorporated into mainline Linux since the first version of the Unbreakable Enterprise Kernel. Please see the initial Oracle Unbreakable Enterprise Kernel Release 2 Release Notes for a detailed description of these changes.

The 2.6.39-300 release is a quarterly driver update release which also includes bug and security fixes.

Note: the actual version number displayed by the kernel and on the RPM packages is 2.6.39. This was done to avoid potential breakage of certain low-level utilities of the Oracle Linux distribution (also known as the "plumbing") that potentially can't cope with the new 3.x version scheme. Regular Linux applications are usually not aware or affected by Linux kernel version numbers.

Notable changes

Driver Updates

The Unbreakable Enterprise Kernel supports a wide range of hardware and devices. In close cooperation with hardware and storage vendors, several device drivers have been updated by Oracle.

Storage drivers
Network drivers

Notable bug fixes

Technology Preview

Following features included in the Unbreakable Enterprise Kernel Release 2 are still under development, but are already made available for testing/evaluation purposes.

Compatibility

Oracle Linux maintains user-space compatibility with Red Hat Enterprise Linux, which is independent of the kernel version running underneath the operating system. Existing applications will continue to run unmodified on Unbreakable Enterprise Kernel Release 2 and no re-certifications are needed for RHEL certified applications.

The Oracle Linux team works closely with 3rd party hardware and software vendors to minimize impact on interoperability during releases but in order to introduce new drivers there are instances where changes must be made. In this release, there are changes to the kernel ABI which requires 3rd party kernel modules on the system be recompiled. Before installing this update, please verify the support status of this release with your application vendor(s).

Known issues

Technology Preview Features for Oracle Linux 5

The following Technology Preview features are currently not supported under Oracle Linux 5 and may not be functionally complete.

These features are not suitable for production use. However, these features are included to provide the feature with wider exposure.

Known Issues for Oracle Linux 5.9

  1. Red Hat compatible Kernel with Oracle bug fixes will be updated to standard Red Hat compatible kernel during ISO upgrade
  2. The Oracle Linux ISO has the patched Red Hat compatible kernel but it is not in the iso repodata, therefore if the ISO is used to upgrade, the kernel will be updated to the Red Hat compatible version of the latest release (the non-patched). This is expected behavior.

  3. Red Hat compatible Kernel will be updated to standard Red Hat compatible kernel with Oracle bug fixes during ULN upgrade
  4. If a 5.8 system running the Red Hat compatible kernel registers to the 5.9 ULN channel, the kernel will be updated to the Red Hat compatible with Oracle bug fixes version of the latest release. This is expected behavior.

  5. Error: "qemudStaretup:1728: Unable to create cgroup for driver: No such device or address"
  6. Use the following steps to workaround this error:

    mkdir /dev/cgroup
    mount -t cgroup none /dev/cgroup -o cpu, memory, devices
  7. Unable to mount btrfs on Oracle Linux 5, error:
  8. mount /dev/sda /mnt
    mount: you must specify the filesystem type
    To correct, add "-t btrfs" to the mount statement or add btrfs in /etc/filesystems

  9. On Oracle Linux 5, the "cp" command did not support reflinks on btrfs
  10. Support for reflink was added to btrfs-prog and can be enabled using "--reflink" as an option on "cp"

  11. YUM command error occurs after system is removed from ULN
  12. After removing a system from the list in ULN and running 'yum update' on the server console, the following error is displayed:

    AttributeError: 'module' object has no attribute 'RhnServerException'
    To work around this issue remove /etc/sysconfig/rhn/systemid and run 'yum update' again

  13. dmsg error: "rtc_cmos: probe of rtc_cmos failed with error -16. Error: Driver 'pcspkr' is already registered, aborting...kudzu[1373] general protection ip:8056c78 sp:bfed96b0 error:0 in kudzu[8048000+26000]"
  14. This is a known issue that is seen with Oracle Linux 5 32-bit para-virtualized guests created using virt-install"

  15. UDEVD to be updated to use oom_score_adj to disable OOM on 2.6.39-UEK2
  16. This change will set com_score_adj, if available otherwise if this fails, it will fall back to the old com_adj
  17. Installation completes successfully but grub.conf file is empty and system will not successfully boot.
  18. This occurs when trying to instal the bootloader to a cciss device. To correct this, edit the drive order by moving /dev/cciss/c0d0 ahead of /dev/sda

  19. Missing openIB drivers (sdp/ulp) for Infiniband in UEK (R1/R2)
  20. When attempting to load the openIB service the error "Loading OpenIB kernel modules:FATAL: Module ib_sdp not found" is displayed. This is an artifact of the ib_sdp driver from OpenIB not being included in the UEK. If you are using OpenIB, you should use the Red Hat compatible kernel.

  21. Service TUX failed to start on Oracle Linux systems running the UEK
  22. When attempting to start tux service (service tux start) the system reports multiple errors for "No such file or directory" and ends with "TUX not implemented in this kernel" TUX is not shipped as part of the UEK and this is expected behavior.

  23. After running yum update there are multiple errors in the yumupdate.log indicating a failure to read modules kvm.ko, kvm-amd.ko, ksm.ko, kvm-intel,ko with the message "no such file or directory"
  24. The error messages are generated due to bad symbolic links in the Red Hat compatible kernel and are considered harmless.

  25. HVM with PV Drivers Upgrade Issue: Cannot Update Bootloader configuration
  26. When upgrading a VM using the HVM with PV Drivers Domain type on Oracle Enterprise Linux 5.9 from previous versions using the ISO media, the option to upgrade the boot loader configuration may not be available. This is due to a difference in the way the kernel detects the disk devices that prevents it from detecting the existing configuration. There are two ways to work around the problem:

      - Use yum to upgrade the systems to 5.9 instead of an ISO
      - Choose the "create new bootloader" option
     
    After the upgrade, old kernel entries that refer to removed kernel packages may remain in the grub.conf configuration file. Make sure to boot to the 5.9 kernel and remove the invalid entries.

  27. No networking after upgrading an Oracle Linux 5 HVM guest on Oracle VM 2.2.2
  28. Upgrading an existing Oracle Linux 5 Update 6 or Update 7 guest which is running in HVM mode on Oracle VM 2.2.2 may result in the network interface(s) being disabled after the reboot post-upgrade. To resolve this issue, ensure that the network interfaces are set to "paravirtual" in Oracle VM Manager.

    Alternatively, to maintain the emulated devices inside the guest, add xen_emul_unplug=never to the kernel boot line in /boot/grub/grub.conf and reboot the guest.

  29. No networking after upgrading an Oracle Linux 5 HVM with PV Drivers guest on Oracle VM 2.2.2 or 3.0
  30. Upgrading an existing Oracle Linux 5 Update 6 or Update 7 guest which is running in HVM with PV Drivers mode on Oracle VM 2.2.2 or 3.0 may result in the network interface(s) being disabled after the reboot post-upgrade. To resolve this issue, rename the /etc/sysconfig/network-scripts/ifcfg-eth0.bak file to /etc/sysconfig/network-scripts/ifcfg-eth0 and reboot the guest.
  31. Warnings appear in dmesg on first boot
  32. After an initial install of Oracle Linux 5.9, the following warning may appear one or more times in the output of dmesg:
    warning: process `firstboot' used the deprecated sysctl system call with 1.23. 
    
    This warning can be safely ignored.
  33. New multipath settings requires changes to /etc/multipath.conf
  34. Oracle Linux 5.9 includes a new version of device-mapper-multipath. This new version requires manual changes to the /etc/multipath.conf file. In the multipath.conf, the line with prio_callout needs to be replaced by the keyword prio, and a simple string, not an executable pathname, used for the parameter to prio. Valid settings for prio are:
  35. Multipath failure causes flush of all multipath devices
  36. In some multipath failure cases, the multipath maps can get flushed when all paths fail, and multipath -l displays no information on the status of your multipath devices. This has been reported with 2.6.39 UEK kernels using NetApp LUNs, but not when running 2.6.18 series kernels with the same storage LUNs.
  37. Error inserting cxgb3i module (13493043)
  38. iscsi offload via cxgb3i on a server with Chelsio T3 Network Controller will fail with following error if ofa kernel package is installed
    fatal: error inserting cxgb3i - unknown symbol t3_l2e_free
    
    Use following workaround steps to fix the issue:
  39. Slow performance on 10Gb Ethernet Interfaces
  40. The default networking tuning parameters may be too low for some 10Gb Ethernet interface drivers. If you are experiencing less than expected network performance, change the following NIC parameters: To:
  41. dbus-daemon warning in dmesg on 32-bit systems
  42. On 32-bit Oracle Linux systems, you may see the following warning in dmesg:
    warning: `dbus-daemon' uses 32-bit capabilities (legacy support in use)
    This is only a warning and can be safely ignored.
  43. iTCO_wdt warning in dmesg
  44. Some server hardware may not support the Intel TCO watchdog drivers. The following warning may be seen in dmesg:
    iTCO_vendor_support: vendor-support=0 
    iTCO_wdt: Intel TCO WatchDog Timer Driver v1.05 
    iTCO_wdt: failed to reset NO_REBOOT flag, reboot disabled by hardware 
    iTCO_wdt: No card detected 
    intel_rng: FWH not detected
    To disable this TCO WatchDog Timer Driver and suppress this warning at boot time add the line "blacklist iTCO_wdt" to /etc/modprobe.d/blacklist-watchdog".
  45. Unbreakable Enterprise Kernel firmware must be manually installed prior to upgrade
  46. Due to a change in the way firmware is loaded between the Red Hat Compatible Kernel and the Oracle Unbreakable Enterprise Kernel, you may need to manually install the correct UEK kernel firmware package if you want to boot a previous kernel version after upgrading to Oracle Linux 5.9. For example, to install the firmware for kernel-uek-2.6.39-300.10.1, download the correct RPM package from the Unbreakable Linux Network and run the following command:
    # rpm -ivh --force kernel-uek-firmware-2.6.39-300.10.1.el5uek.noarch.rpm
    
  47. Memory limited for Oracle Linux 5.9 Paravirtualized Guests
  48. Oracle Linux 5.9 paravirtualized guests running on Oracle VM with the Unbreakable Enterprise Kernel have a memory limit of 480GB. Setting the memory beyond this value may cause the paravirtualized network interfaces to fail.
  49. Oracle Linux 5.9 32-bit guest in HVM with PV Drivers mode with more than 1 vCPU hangs on boot
  50. Attempting to boot a 32-bit HVM guest with PV Drives with more than 1 virtual CPU may fail. To resolve this issue, boot the guest with a single virtual CPU and append tsc=reliable parameter to the kernel boot line in */boot/grub/grub/conf*. You can then boot the guest with multiple virtual CPUs.
  51. kexec/kdump may not work on some server platforms
  52. Certain server platforms may not support the default kdump parameters. To test this, enable kdump and issue:
    echo c > /proc/sysrq-trigger
    If your server hangs indefinitely, you may need to disable the "reset_devices" parameter for kdump. To do this, edit /etc/sysconfig/kdump and remove the "reset_devices" parameter from the KDUMP_COMMANDLINE_APPEND variable.
  53. Using yum to upgrade previous update from Unbreakable Linux Network (ULN)
  54. yum package need to be updated before upgrading previous update releases of Oracle Linux 5 to Oracle Linux 5 Update 8 using yum from ULN. Please use the following steps to upgrade:
       # yum update yum
       # yum update
    
  55. Default kernel after upgrade
  56. In certain upgrade scenarios the Unbreakable Enterprise Kernel may not be the default boot kernel. Update /etc/grub.conf to make the latest Unbreakable Enterprise Kernel as the default boot kernel.
  57. Missing kernel module ib_iser.ko
  58. Starting OpenFabrics Enterprise Distribution 1.4.2-5 version (ofa package), ib_iser kernel module is not built as part of ofa package. If the ofa package installed on the system tries to load ib_iser.ko, the following message will be displayed on the console
     ib_iser: disagrees about version of symbol ib_fmr_pool_unmap 
     ib_iser: Unknown symbol ib_fmr_pool_unmap 

    These messages can be ignored as ib_iser.ko is no more needed.
  59. Upgrade failure for the full install of Oracle Linux 5
  60. If any of the following packages are installed on the system
    kmod-cmirror
    kmod-cmirror-xen
    kmod-gfs
    kmod-gfs-xen
    kmod-gnbd
    kmod-gnbd-xen
    
    Upgrade of Oracle Linux 5 system may fail with the following error
    An error occurred while installing packages.
    Please examine /root/install.log on your 
    installed system for detailed information
    
    to workaround this problem, remove above packages before upgrade. These packages can be installed after the upgrade.
  61. Failed to stat /dev/mapper/* message during boot of paravirtualized guests
  62. Following messages during boot of paravirtualized guest can be ignored
      Failed to stat /dev/mapper/no
      Failed to stat /dev/mapper/block
      Failed to stat /dev/mapper/devices
      Failed to stat /dev/mapper/found
    
    These errors are expected as dmraid utility returns 'no block devices found' because an ioctl to get serial numbers on pv devices returns -EINVAL."
  63. Spurious udev messages during restore of a failed path (11682171)
  64. When failed paths are restored in a multipath configuration, you may see udevd-work error messages in /var/log/messages. The failed paths do get restored despite these messages.
  65. Default NFS v4 mount options
  66. The default NFS mount option has changed to NFS v4. To mount an NFS volume using NFS v3 (the default using the Red Hat Compatible Kernel), use the following mount options:
    -o vers=3,mountproto=tcp
    
  67. Serial console not working for a paravirtualized guest (10059184)
  68. Please use the following steps to configure the serial console in a paravirtualized guest (PV) using the Unbreakable Enterprise kernel (updates are in the guest):
    1. change the kernel boot parameter 'console=xvc0' to 'console=hvc0' in the /etc/grub.conf:
    2.     kernel /vmlinuz-2.6.32-300.10.1.el5 ro root=/dev/sd0 console=tty0 console=hvc0
      
    3. add an 'hvc0' entry to /etc/inittab:
    4.     co:2345:respawn:/sbin/agetty hvc0 9600 vt100-nav
      
    5. add an 'hvc0' entry to /etc/securetty
    .
  69. Setting serial console in a hardware virtualized (HVM) guest
  70. To set the serial console a hardware virtualized (HVM) guest use the following settings in the guest:
  71. Console appears to hang while booting on certain systems (10094052)
  72. On some hardware, the console may appear to hang during the boot process after starting udev. But the system does boot up properly and is accessible. A workaround to this problem is to add nomodeset as a kernel boot parameter in /etc/grub.conf:
        kernel /vmlinuz-2.6.32-300.10.1.el5uek.x86_64 ro root=/dev/sd0 nomodeset
    
  73. Unable to start X (10056679)
  74. On some hardware, startup of the X environment may fail after upgrading to Unbreakable Enterprise Kernel with following Error:
        Fatal server error:
        Cannot run in framebuffer mode. Please specify busIDs for all
        framebuffer devices.
    
    To workaround this issue reconfigure X using command "X -configure" and use generated configuration file to start X environment.
  75. Higher memory is required for crashkernel option
  76. With the Unbreakable Enterprise Kernel a higher memory setting is required for crashkernel option in /etc/grub.conf file. Booting with crashkernel=128M@16M will result in following error:
        crashkernel reservation failed - memory is in use
    
    Minimum value for crashkernel option is 128M@32M. Based on the system configuration a higher value may be required. If kdump service loading fails increase the second value (32M). If the system hangs or crashes with out of memory during dumping core increase the first value (128M)
  77. High IO loads on NFS using mmap may fail (11071875)
  78. In some instances, large IO loads on NFS that use mmap may fail and cause the kernel to panic with a message like:
        kernel BUG at mm/truncate.c:469!
    
  79. NFS: Error message displayed while mounting an NFS filesystem (10070721)
  80. The following error may be encountered while mounting an NFS filesystem:
        "SVC: FAILED TO REGISTER LOCKDV1 RPC SERVICE (ERRNO 97)."
    
    This is an informative message and can be ignored
  81. Unable to remove directory during boot
  82. The following message during boot can be ignored as SELinux does not allow to recreate the login subdirectory and hence directory should not be removed for dovecot:
    Error :
      "rm: cannot remove `/var/run/dovecot/login': Is a directory "
    
  83. Upgrade ocfs2 and oracleasm kernel modules
  84. ocfs2 and oracleasm kernel modules will not be automatically updated during upgrade of Oracle Linux 5 as package name for these modules contains kernel version in it. These packages need to be updated manually after the upgrade.
  85. device-mapper-multipath package for Red Hat compatible kernel
  86. Unbreakable Enterprise Kernel requires kpartx and device-mapper-multipath RPMs that are newer than those provided for the Red Hat compatible kernel. By default device-mapper-multipath-0.4.9-56.0.3.el5 and kpartx-0.4.9-56.0.3.el5 are installed. These updated versions of kpartx and device-mapper-multipath are supported with the Red Hat compatible kernel. The original RPMs for the Red Hat compatible kernel (kpartx-0.4.7-54.el5 and device-mapper-multipath-0.4.7-54.el5) are also included on the Oracle Linux 5.9 ISO in directory Server/oracle_updated.


  87. Workaround to use pirut and system-config-packages to manage packages
  88. Pirut and system-config-packages cannot be used to manage, install, and uninstall software packages. Follow the steps below to resolve the issue

  89. Conflicting packages
  90. The following packages can not co-exists due to file conflict

  91. Unresolved dependency during update using up2date
  92. If updating from Oracle Linux 5 GA or RHEL 5 GA using up2date, you may encounter the following error:
      Unresolvable chain of dependencies:
      xen-3.0.3-64.el5    requires xen-hypervisor-abi = 3.1
    
    This happens because a kernel-xen package update is needed by the xen package, but 'kernel*' is part of the up2date pkgSkipList. To get around this, add 'xen;xen-devel;' to the pkgSkipList line in /etc/sysconfig/rhn/up2date and then run up2date. You can update the xen and kernel-xen packages together later via 'up2date --force xen kernel-xen'.
  93. /ETC/MULTIPATH/WWIDS is missing after multipath-root install on OL5.9
  94. When installing onto a multipath root disk, the file /etc/multipath/wwids is not automatically generated at boot time, because the root disk (which is itself on multipath) is still in read-only mode at the time multipath services start up. This will result in a warning message displayed to the console at boot time:
    Cannot open file [/etc/multipath/wwids] readonly: No such file or directory

    This message is a warning that can be ignored safely; it does not affect the runtime operation of multipathd or device-mapper-multipath. To squelch this warning message, after the system has booted, it is possible to generate the missing wwids file with this command (executed as root):
    multipath -l

    A future errata update should include a complete fix for this issue.

Configuring Updates for Oracle Linux 5 Update 9

Unbreakable Linux Network

Oracle Linux offers an option to keep your operating system up to date with latest operating system patches using up2date or yum.

To register with ULN, use the following command:

# up2date --register

To access Linux updates via Unbreakable Linux Network, you must purchase a Linux support subscription. For more information please visit http://linux.oracle.com.

During ULN registration the server will be automatically registered with two channels: the latest channel for the base repository as well as the latest channel for the Unbreakable Enterprise Kernel Release 2. If you have upgraded from a previous release and do not wish for the UEKR2 to be installed, you will need to manually unsubscribe the server from this channel.

The Unbreakable Linux Network includes Oracle-specific software packages, for example Oracle's ASMlib userspace package and the Oracle Instant Client. To enable access to these packages, login to the Unbreakable Linux Network and subscribe to the "Oracle Software" channel.

Note: Prior to using up2date, import the RPM-GPG-KEY for Oracle Linux:

# rpm --import /etc/pki/rpm-gpg/RPM-GPG-KEY

Public Yum

Oracle now provides all errata and updates for Oracle Linux via the public yum service. This service does not require a Linux support subscription, but only includes updates to the base distribution and does not include Oracle-specific software. To enable updates via public-yum, please visit http://public-yum.oracle.com and follow the instructions on that website.

By default, all new installs of Oracle Linux 5 Update 9 are automatically configured to use the public-yum update service. No modification is required to use this service. The public-yum service is automatically disabled when a server is registered with the Unbreakable Linux Network.

Installation Media

Note: Oracle Linux 5 Update 9 now contains two distinct repository sources on the installation media. To configure yum to use the installation media as a yum source, create the following file /etc/yum.repos.d/Media.repo with content similar to the following:

[ol5_base_media]
name=Oracle Linux 5 Update 9 Base Media
baseurl=file:///media/Server
gpgcheck=0
enabled=1

[ol5_uek_media]
name=Oracle Linux 5 Update 9 UEK Media
baseurl=file:///media/UEK2
gpgcheck=0
enabled=1

You may need to adjust the baseurl parameter to match where you mount the ISO media on your server.