=== tgm4883_ is now known as tgm4883 === monk is now known as Guest63787 === ulkesh_ is now known as ulkesh [08:48] hello when i dump blocked tasks via sysrq (echo w> /proc/sysrq-trigger) my output is dominated by schedstats and this overruns the kmsg buffer and i cannot see the blocked stasks, how can i disable this? [08:48] i'm using 4.4.0-62-generic kernel === caribou_ is now known as caribou === shadeslayer_ is now known as shadeslayer === mhcerri_ is now known as mhcerri [12:34] rtg: my patch works, in addition to building. I am going to update all the drivers today [12:35] tseliot, great, I'll get that 4.10 kernel promoted to -proposed [12:35] rtg: I can't test CPU_HOTPLUG with my hardware though [12:35] tseliot, well, I guess we'll have to rely on community [12:37] rtg: fair enough, it's a bit of a corner case anyway [12:53] My system is crashed...Who can help me please? http://paste.ubuntu.com/24000622/ [12:54] now I run "crash" now... [12:55] It;s openstack hypervisor ..and 5 machines are crashed in parallel [12:57] osc_khoj, that is reporting that the kernel was hit with an unexpected NMI and got upset [12:58] awp : but hp enginner said there are no log in ilo. [12:58] A Non-Maskable Interrupt (NMI) is a hardware interrupt that cannot be ignored by the processor. These types of interrupts are usually reserved for very important tasks and to report hardware errors to the processor. The Operating system does not have much of a role in NMI handling, it primarily reports what the hardware throws. The kernel is able to interpret the reason for some NMIs. The normal behaviour is to log it an [12:59] I searched the google..but, I need some clue to solve the problem.thanks.apw [13:00] osc_khoj, as that text says, the kernel doesn't normally have anything to do with them, they come from the platform [13:00] osc_khoj, i would tend to expect that to be a hardware problem if it occured on my machine. i would likely reboot as normal and worry about it if it recurs [13:00] (on my own kit, you obviously need to decide for your platform) [13:02] Is there any command to find the clue in crash? HW : HP BL460c G9 - OS : ubuntu 14.04 (3.16.0-30-generic #40~14.04.1-Ubuntu) - SW : openstack Juno version [13:03] after crashed, I update kernel to 3.16.0-77..because there are no debug tool...in 3.16.0-30. [13:04] but, same issue occured.. [13:07] osc_khoj, on the same CPU do you know ? [13:08] osc_khoj, that crash is pretty thin on information if all it is showing is that one processor in NMI [13:08] apw, many machine occured the crash. [13:09] we have 20 openstack compute node...5 machines crashed about 1 month... [13:09] I run some command if you want, apw, now [13:11] osc_khoj, hmmm. 3.16 that was a utopic kernel, so we don't have anything newer on that stream to try, it being far beyond EOL [13:12] osc_khoj, it must have any number of guest to host security issues [13:12] i would try a kernel from a supprted release on there as the first step, to see if that stops the issue [13:13] It's very difficult to update the kernel because it;s openstack compute node.. [13:14] Install kernel, migrate guests off, reboot, repeat? [13:14] osc_khoj, really, don't you just migrate them to another node reboot it, and migrate them back ? [13:17] hmm,after update the kenrel, I can run live-migrate and can test...It don;t need to com back. [13:18] but we want to know the core issue...by analyze the crash dump. [13:18] Is there no info in crash? ^^ [13:19] osc_khoj, maybe, likely not, it looks and sounds like the h/w is trying to tell you something and the kernel has no idea how to handle it and exploding [13:20] osc_khoj, that is the kind of thing we see new drivers for, handling new forms of EDAC information coming up from the hardware [13:23] Thanks apw, I will update the kernel and retest h/w.... === alai` is now known as alai [16:14] I can't kill -9 a process running in a docker container (tried both from the host and from the container). ps status is R. Any idea? === JanC_ is now known as JanC [17:19] hi [17:21] I am looking for good instructions for recompiling 4.8.0-28-generic kernel. I need to edit a file, recompile. I'd need that specific kernel as it works with sound on a new machine [17:22] I haven't compiled kernels before so beginner type instructions would be useful and the location of source for download [18:24] green_, there should be information in the kernel team wiki [18:24] https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel perhaps [18:38] https://bpaste.net/show/1b11450f5e8e - 'make mrproper' cleans everything including delete debian/rules which is needed for compilation [18:40] WeiJunLi, yes, it has no understanding of the debian bits [18:41] apw: what that means exactly [18:41] what should i do [18:41] running make mrproper does not work correctly in an ubuntu kernel, but it should not be needed [18:41] to clean up: fakeroot debian/rules clean [18:42] I have done that. [18:42] and then fakeroot debian/rules editconfigs [18:42] and for last that one in the paste [18:42] which productes that Error [18:43] produces* [18:44] WeiJunLi, rmdir include/config [18:45] apw: I dont have any file named config inside include folder [18:45] I do have one .config which in the root of the source tree [18:45] which is the kernel .config itself [18:46] i find that when it produces that error rmdir include/config sorts that out [18:46] 'no such file or directory' [18:47] apw: ok thanks. If I recompile would the resulting kernel be exactly the same as the one I am using exceept for the changes I make? === jdstrand_ is now known as jdstrand [18:51] apw: anything else? [18:52] WeiJunLi, it is possible the .config would also confuse things [18:52] WeiJunLi, you don't normally use a .config in the top level with debian packaging [18:53] I just noticed now that I have a bunch of .deb files on /home/ [18:54] not sure if I should go forward or just delete them [18:56] apw fakeroot debian/rules editconfigs [18:56] I'm loading a .config [18:57] not sure what's wrong. [18:57] i am not sure what you are askingme [18:58] where should the .config be then [18:58] if not in the top level [18:58] in debian/configs is where our configs are [18:58] ok so I should move the .config there [18:58] and what commands to do next [18:59] if you are trying to make an ubuntu compatible kernel it is more normal [18:59] to add the changes you need to make to the debian/configs/.../config.flavour. file [19:00] and then run fakeroot debian/rules updateconfigs [19:00] replacing the .config wholesale is not a normal thing to do [19:00] but if you really think you have to dropping it in as that leaf config and running the same command may help [19:02] apw: I'm used with the basic method, make; make modules; sudo make modules_install; make install [19:03] everytime i say I do that way, the feedback is to use the debian way with fakeroot [19:03] but in fact in kernel # nobody do that wya. [19:03] you use whatever method works for you [19:04] apw: Im trying fakeroot because last time i compiled one, guess was missing the filesystem '/' since when the kernel booted i was getting into the busybox initramfs shell [19:05] or your config doesn't contain something that the ubuntu runtime initramfs requires [19:10] apw: if i show you my .config can you try to spot something wrong? [19:10] WeiJunLi, it is far to complex to just look over the config [19:11] WeiJunLi, you would want to compare the generated config, with the config you intend [19:11] WeiJunLi, you can generate configs using "fakeroot debian/rules genconfigs" ... they will be generated in CONFIGS/* === manjo` is now known as manjo === cjwatson_ is now known as cjwatson === mwhudson_ is now known as mwhudson === mwhudson is now known as Guest32650 === mwhudson_ is now known as mwhudson