I'm running our test suite as part of pre-release checks and I get
constant failures of libcgrouptest01.
The test is quite confusing to me, but IMHO does not do anything harmful
and works pretty well on Fedora 14.
However, on current rawhide something kills the tests when it reads
/dev/cgroup_controllers-1/group_a/memory.soft_limit_in_bytes:
...
stat("/dev/cgroup_controllers-1/group_a/memory.soft_limit_in_bytes",
{st_mode=S_IFREG|0644, st_size=0, ...}) = 0
open("/dev/cgroup_controllers-1/group_a/memory.soft_limit_in_bytes",
O_RDONLY|O_CLOEXEC) = 4
fstat(4, {st_mode=S_IFREG|0644, st_size=0, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
= 0x7f5c23b02000
read(4, <unfinished ...>
+++ killed by SIGKILL +++
I turned off SELinux, the first obvious cause, but without any effect.
Were there any changes in the kernel which could lead to such kill? Can
I track source of the kill signal somehow?
Simple cat works like charm:
open("/dev/cgroup_controllers-1/group_a/memory.soft_limit_in_bytes",
O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=0, ...}) = 0
fadvise64(3, 0, 0, POSIX_FADV_SEQUENTIAL) = 0
read(3, "9223372036854775807\n", 32768) = 20
write(1, "9223372036854775807\n", 20) = 20
Jan
------------------------------------------------------------------------------
Try before you buy = See our experts in action!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-dev2
_______________________________________________
Libcg-devel mailing list
Libcg-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libcg-devel
