Testing the new 2.4.26 -om-2 patch with Migshm and
flush() enabled results in a reproduceable kernel
panic; Kernel BUG at page_alloc.c:98!.
After rebooting the cluster to the oM kernel, I ran a
test compile of the kernel with distcc, at about 3
minutes 9 of the fifteen backend nodes paniced
togther, all with the same error. the compile of course
errored out. I left the cluster as it was, with 6
backend and home node running. Come back in to work
this morning to find the remaing 6 backend nodes
paniced. Only the home node was left up and running.
#
# openMosix
#
CONFIG_MOSIX=y
# CONFIG_MOSIX_TOPOLOGY is not set
CONFIG_MOSIX_SECUREPORTS=y
CONFIG_MOSIX_DISCLOSURE=3
CONFIG_MOSIX_PIPE_EXCEPTIONS=y
CONFIG_MOSIX_NO_OOM=y
# CONFIG_MOSIX_EXT_LOCALTIME is not set
Also I just noticed that the migshm,flush() and debug
Configs are not listed in the config file. I am sure
that I have migshm compiled in becuase it is listed as
starting on kernel boot.
Logged In: YES
user_id=1021609
Update- I can reproduce the exact same thing on Kris's rpms
for 2.4.26-om-2
Logged In: YES
user_id=1021609
And I managed to produce the same result with the 2.4.26-om1
prebuilt RPMs
Logged In: YES
user_id=1283799
Originator: NO
2.4 bug