Difference between revisions of "Download/kernel/rhel5/028stab067.4/changes"

Revision as of 18:20, 26 January 2010

Rebase to 2.6.18-164.10.1.el5 kernel: a number of security and bug fixes
Bug fixes for the following issues:
- A Hardware Node may hang (with "Route hash chain too long! Adjust your secret_interval!" messages) due to a bug in the on demand emergency route cache flushing code (OpenVZ Bug #1378)
- The values of entries in the /proc/diskstats file may be set to zero (OpenVZ Bug #1373)
- A kernel panic may occur on an attempt to mount a corrupted partition if the 'reiserfs' module is loaded.
- A kernel bug may occur due to the timer list corruption if several Containers are using NFS mounts at the same time
- A kernel panic may occur when deleting an audit rule from the Hardware Node (OpenVZ Bug #1351)
- A kernel panic may occur on the Destination Node during the online migration of a Container with a TUN device
- A kernel panic occurs on loading the DRBD module (OpenVZ Bug #1371)
The load balancing mechanism in the CPU fair scheduler has been improved, especially the distribution of the CPU time on SMP systems with a small number of processes

@@ Line 4: / Line 4: @@
 * Rebase to 2.6.18-164.10.1.el5 kernel: a number of security and bug fixes
 * Bug fixes for the following issues:
-** A Hardware Node may hang due to a bug in the on demand emergency route cache flushing code. The following messages are written to the log file before the Node hangs: "Route hash chain too long! Adjust your secret_interval!" ({{Bug|1378}})
+** A Hardware Node may hang (with "Route hash chain too long! Adjust your secret_interval!" messages) due to a bug in the on demand emergency route cache flushing code ({{Bug|1378}})
 ** The values of entries in the /proc/diskstats file may be set to zero ({{Bug|1373}})
 ** A kernel panic may occur on an attempt to mount a corrupted partition if the 'reiserfs' module is loaded.