We have just released a new RHEL6-based kernel, 042test005. It is shaping up pretty good — as you can see from the changelog, it's not just bug fixes but also performance improvements. If you haven't tried it yet, I suggest to do it today! Do not postpone this until 2011 — after all, this is what will become the next stable OpenVZ kernel.
RHEL6 kernel needs an appropriate (i.e. recent) Linux distribution. If you don't want latest Fedora releases, can't afford RHEL6, and tired of waiting for CentOS 6, I suggest you go with Scientific Linux 6 (SL6). This is yet another RHEL6 clone developed and used by CERN, Fermilabs and other similar institutions.
While SL6 is still at its infancy (they have recently released alpha 3 and plan to release beta 1 at Jan 7 2011), it it worth trying since it's based on a very stable set of sources from RHEL6. Repositories and stuff are available from http://ftp.scientificlinux.org/linux/scientific/6rolling/
RHEL6 kernel needs an appropriate (i.e. recent) Linux distribution. If you don't want latest Fedora releases, can't afford RHEL6, and tired of waiting for CentOS 6, I suggest you go with Scientific Linux 6 (SL6). This is yet another RHEL6 clone developed and used by CERN, Fermilabs and other similar institutions.
While SL6 is still at its infancy (they have recently released alpha 3 and plan to release beta 1 at Jan 7 2011), it it worth trying since it's based on a very stable set of sources from RHEL6. Repositories and stuff are available from http://ftp.scientificlinux.org/linux/scientific/6rolling/


Comments
Please remove unused put_fs_struct symbol from kernel/exit.c
;)
This trash is defined under #ifdef CONFIG_MM_OWNER, which is not compatible with our kernel =)
+EXPORT_SYMBOL_GPL(put_fs_struct);
It is in your patchset. I am using CONFIG_MM_OWNER because it is needed by dm-ioband cgroup patchset. I am going to use it and make few test with your great kernel.
Blkio throttle V1 patchset by Vivek working great but doing kernel panic on scsci subsystem randomly... (and only for software raid arrays... ;))
Would be really great if virtuozzo team will merge and open the iolimit functionality from virtuozzo kernel to openvz.
That's why I am working on it to have the iolimit throttle for openvz.
I need to give Scientific Linux a try. I have played with it a little in the past but unlike CentOS, they seem to stray a bit from the stock Red Hat packages. The CentOS developers are pretty hush hush and have not given any indicator as to when to expect the CentOS 6 release. My guess would be sometime in February.
also is the vswap a "overlay" for UBC or is it a whole new system for resources ?
So will vswap set the UBC settings for you or is it a new system ?
Hmm, looks like it's for you only, ie I haven't seen any other report like this one.
Care to get some logs and file a bug to http://bugzilla.openvz.org/?
here it says that the kernel has native cgroup support (http://wiki.openvz.org/Download/kernel/rhel6/042test003.1). I tried to use it with cpu cgroup limiting (used cpu.rt_period_us and cpu.rt_runtime_us for hard CPU limits) but it doesn't work.
Kir, did you tried it or any ETA when this will work ? OpenVZ kernel create the VID hierarchy tree with cgroup objects but as I wrote It doesn't work.
root@kvm01:/cgroup/cpu/101# ls
cgroup.procs cpu.rt_period_us cpu.rt_runtime_us cpu.shares notify_on_release self_destruction tasks
Please file a bug to http://bugzilla.openvz.org/ — this is the best way of saying to developers “guys, please take a look”