Known Issues
The following known issues are for the Red Hat-compatible
kernels for Oracle Linux 5 Update 10 and Oracle Linux 6 Update 5.
|
|
1948587 |
On RHEL 6 Update 5 Hosts, Host Drivers Cannot Load When Using a
Standard
initd/vmlinuz
Image for PXE Boot
If you use the vmlinuz and initrd.img from
the RHEL ISO, a problem prevents the vNICs and vHBAs from being
accessible even though they are injected in to the custom boot image. Consequently,
the RHEL 6 Update 5 host will not PXE or SAN Boot.
Workaround: For SAN or PXE Boot of RHEL 6 Update 5 hosts
over an Oracle Fabric Interconnect, you must always use the
vmlinuz and initrd.img
from Oracle Linux instead of the Red Hat Enterprise Linux versions
of these tools. |
19455744 |
On RHEL 6 Update
5 hosts, SAN Booting Causes Kernel Panic When Accessing Boot Image
While creating a SAN boot image, an internal script inserts the Oracle Virtual Networking host
drivers into initrd.img file instead of the
initramfs.img that the REHL 6 Update 5 host
expects. Although SAN install works correctly, when the server
attempts to boot up, it experiences a kernel panic with an
Init Not Tainted
message.
Workaround: Run the following command
after SAN install but
before rebooting the server:
/sbin/mkinitrd -f --preload mlx4_ib --preload mlx4_core --preload xscore --preload xsvhba --preload xsvnic --with dm-multipath /boot/initramfs-2.6.32-431.el6.x86_64.img 2.6.32-431.el6.x86_64 |
19380352 |
During Reboot, an Oracle Linux 5 Update 10 Server With Multiple vNICs
Experiences a Kernel Panic and Crashes
If you reboot an Oracle Linux 5 Update 10 host running 5.0.8
host drivers with two or more vNICs configured on it, the host will
experience a kernel panic, then crash.
Workaround: Stop the openibd service
before rebooting the server with two or more vNICs:
Stop the xsigo service: service xsigo stop
Stop the openibd service: service openibd stop
Deactivate the openibd service
so that it will not start again after server reboot: chkconfig openibd off
Start the xsigo service: service xsigo start
|
16862166 |
Server Profile Staysup/down
During PXE Boot on Oracle
Linux
During a PXE boot image installation, the server profile on
the Oracle Fabric Interconnect might stay in up/down state.
After the host reboots and the host driver is loaded into memory,
the server profile successfully comes online in an up/up state.
Workaround: No workaround exists
at this time. The PXE installation is successful. |
16335100 |
Seven-Second
Multicast Traffic Hit When Subnet Manager Is Started
If an event causes a failover or restart of the OpenSM InfiniBand
Subnet Manager, multicast traffic is delayed for approximately seven
seconds while the failover or restart occurs.
Workaround: No workaround exists
at this time. |
|