intel / numatop

NumaTOP is an observation tool for runtime memory locality characterization and analysis of processes and threads running on a NUMA system.
BSD 3-Clause "New" or "Revised" License
197 stars 47 forks source link

NumaTOP failed to start #19

Closed hghorayeb closed 6 years ago

hghorayeb commented 8 years ago

Hello,

I am using the last version of numatop code, i successfully run make install. however i am facing a blocking issue to start numatop. here below the details:

Please note that I didn't apply the kernel patch you mentioned. I didn't succeed to apply it, maybe a hint on the linux command to apply this patch could help.

bash$ numatop NumaTOP is starting ... Fail to setup perf (probably permission denied)!

with root: bash$ numatop NumaTOP is starting ... Segmentation fault (core dumped) ibm001srv root /nfshome/nonuser/autoengine/

Operating Environment: bash$ uname -a Linux ibm001srv 4.5.4-1.el7.elrepo.x86_64 #1 SMP Thu May 12 12:17:54 EDT 2016 x86_64 x86_64 x86_64 GNU/Linux

bash$ which numatop /usr/local/bin/numatop ibm001srv autoengine ~/ bash$ ls -l /usr/local/bin/numatop -rwxr-xr-x. 1 root root 541174 May 16 17:07 /usr/local/bin/numatop*

bash$ cat /proc/cpuinfo | grep 'model name' | uniq model name : Intel(R) Xeon(R) CPU E5-2690 0 @ 2.90GHz

yaoj commented 8 years ago

Please remove the top patch “2327110 Add cache monitoring and memory bandwidth monitoring features”, because it strongly relies on the kernel patch (CQM/MBM patch) which I mentioned. If you didn’t apply the kernel patch, there would be the issue.

I will provide a patch some time later for getting better compatibility on the kernel which is not applied with the required CQM/MBM kernel patch.

Thanks Jin Yao

From: Hicham GHORAYEB [mailto:notifications@github.com] Sent: Monday, May 16, 2016 11:28 PM To: 01org/numatop Subject: [01org/numatop] NumaTOP failed to start (#19)

Hello,

I am using the last version of numatop code, i successfully run make install. however i am facing a blocking issue to start numatop. here below the details:

Please note that I didn't apply the kernel patch you mentioned. I didn't succeed to apply it, maybe a hint on the linux command to apply this patch could help.

bash$ numatop NumaTOP is starting ... Fail to setup perf (probably permission denied)!

with root: bash$ numatop NumaTOP is starting ... Segmentation fault (core dumped) ibm001srv root /nfshome/nonuser/autoengine/

Operating Environment: bash$ uname -a Linux ibm001srv 4.5.4-1.el7.elrepo.x86_64 #1https://github.com/01org/numatop/pull/1 SMP Thu May 12 12:17:54 EDT 2016 x86_64 x86_64 x86_64 GNU/Linux

bash$ which numatop /usr/local/bin/numatop ibm001srv autoengine ~/ bash$ ls -l /usr/local/bin/numatop -rwxr-xr-x. 1 root root 541174 May 16 17:07 /usr/local/bin/numatop*

bash$ cat /proc/cpuinfo | grep 'model name' | uniq model name : Intel(R) Xeon(R) CPU E5-2690 0 @ 2.90GHz

— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHubhttps://github.com/01org/numatop/issues/19

hghorayeb commented 8 years ago

I removed the patch:

bash$ git fetch origin 92cbe59ff3abc74356801b471b281f03443a3e23 From https://github.com/01org/numatop

and successfully built the numatop: bash$ make gcc -g -Wall -O2 -o cmd.o -c common/cmd.c gcc -g -Wall -O2 -o disp.o -c common/disp.c gcc -g -Wall -O2 -o lwp.o -c common/lwp.c gcc -g -Wall -O2 -o numatop.o -c common/numatop.c gcc -g -Wall -O2 -o page.o -c common/page.c gcc -g -Wall -O2 -o perf.o -c common/perf.c gcc -g -Wall -O2 -o proc.o -c common/proc.c gcc -g -Wall -O2 -o reg.o -c common/reg.c gcc -g -Wall -O2 -o util.o -c common/util.c gcc -g -Wall -O2 -o win.o -c common/win.c gcc -g -Wall -O2 -o os_cmd.o -c common/os/os_cmd.c gcc -g -Wall -O2 -o os_perf.o -c common/os/os_perf.c gcc -g -Wall -O2 -o os_win.o -c common/os/os_win.c gcc -g -Wall -O2 -o node.o -c common/os/node.c gcc -g -Wall -O2 -o map.o -c common/os/map.c gcc -g -Wall -O2 -o os_util.o -c common/os/os_util.c gcc -g -Wall -O2 -o plat.o -c common/os/plat.c gcc -g -Wall -O2 -o pfwrapper.o -c common/os/pfwrapper.c gcc -g -Wall -O2 -o sym.o -c common/os/sym.c gcc -g -Wall -O2 -o os_page.o -c common/os/os_page.c gcc -g -Wall -O2 -o wsm.o -c intel/wsm.c gcc -g -Wall -O2 -o snb.o -c intel/snb.c gcc -g -Wall -O2 -o nhm.o -c intel/nhm.c gcc -g -Wall -O2 -o bdw.o -c intel/bdw.c gcc -g -o numatop cmd.o disp.o lwp.o numatop.o page.o perf.o proc.o reg.o util.o win.o os_cmd.o os_perf.o os_win.o node.o map.o os_util.o plat.o pfwrapper.o sym .o os_page.o wsm.o snb.o nhm.o bdw.o -lncurses -lpthread -lnuma

however, I still have the issue:

bash$ ./numatop NumaTOP is starting ... Fail to setup perf (probably permission denied)!

yaoj commented 8 years ago

You have to run numatop by root.

From: Hicham GHORAYEB [mailto:notifications@github.com] Sent: Tuesday, May 17, 2016 8:16 PM To: 01org/numatop Cc: Jin, Yao; Comment Subject: Re: [01org/numatop] NumaTOP failed to start (#19)

I removed the patch:

bash$ git fetch origin 92cbe59https://github.com/01org/numatop/commit/92cbe59ff3abc74356801b471b281f03443a3e23 From https://github.com/01org/numatop

and successfully built the numatop: bash$ make gcc -g -Wall -O2 -o cmd.o -c common/cmd.c gcc -g -Wall -O2 -o disp.o -c common/disp.c gcc -g -Wall -O2 -o lwp.o -c common/lwp.c gcc -g -Wall -O2 -o numatop.o -c common/numatop.c gcc -g -Wall -O2 -o page.o -c common/page.c gcc -g -Wall -O2 -o perf.o -c common/perf.c gcc -g -Wall -O2 -o proc.o -c common/proc.c gcc -g -Wall -O2 -o reg.o -c common/reg.c gcc -g -Wall -O2 -o util.o -c common/util.c gcc -g -Wall -O2 -o win.o -c common/win.c gcc -g -Wall -O2 -o os_cmd.o -c common/os/os_cmd.c gcc -g -Wall -O2 -o os_perf.o -c common/os/os_perf.c gcc -g -Wall -O2 -o os_win.o -c common/os/os_win.c gcc -g -Wall -O2 -o node.o -c common/os/node.c gcc -g -Wall -O2 -o map.o -c common/os/map.c gcc -g -Wall -O2 -o os_util.o -c common/os/os_util.c gcc -g -Wall -O2 -o plat.o -c common/os/plat.c gcc -g -Wall -O2 -o pfwrapper.o -c common/os/pfwrapper.c gcc -g -Wall -O2 -o sym.o -c common/os/sym.c gcc -g -Wall -O2 -o os_page.o -c common/os/os_page.c gcc -g -Wall -O2 -o wsm.o -c intel/wsm.c gcc -g -Wall -O2 -o snb.o -c intel/snb.c gcc -g -Wall -O2 -o nhm.o -c intel/nhm.c gcc -g -Wall -O2 -o bdw.o -c intel/bdw.c gcc -g -o numatop cmd.o disp.o lwp.o numatop.o page.o perf.o proc.o reg.o util.o win.o os_cmd.o os_perf.o os_win.o node.o map.o os_util.o plat.o pfwrapper.o sym .o os_page.o wsm.o snb.o nhm.o bdw.o -lncurses -lpthread -lnuma

however, I still have the issue:

bash$ ./numatop NumaTOP is starting ... Fail to setup perf (probably permission denied)!

— You are receiving this because you commented. Reply to this email directly or view it on GitHubhttps://github.com/01org/numatop/issues/19#issuecomment-219700059

hghorayeb commented 8 years ago

With the root there is a core dumped:

ibm001srv root /nfshome/nonuser/autoengine/ bash$ numatop NumaTOP is starting ... Segmentation fault (core dumped)

bash$ ldd /usr/local/bin/numatop linux-vdso.so.1 => (0x00007ffcbf576000) libncurses.so.5 => /lib64/libncurses.so.5 (0x00007fe53dce3000) libtinfo.so.5 => /lib64/libtinfo.so.5 (0x00007fe53dab9000) libpthread.so.0 => /lib64/libpthread.so.0 (0x00007fe53d89c000) libnuma.so.1 => /lib64/libnuma.so.1 (0x00007fe53d690000) libc.so.6 => /lib64/libc.so.6 (0x00007fe53d2cf000) libdl.so.2 => /lib64/libdl.so.2 (0x00007fe53d0ca000) /lib64/ld-linux-x86-64.so.2 (0x000055705003f000) libgcc_s.so.1 => /lib64/libgcc_s.so.1 (0x00007fe53ceb4000)

yaoj commented 8 years ago

No issue occurred at my side. Could you send me with the core file in another mail thread? yao.jin@intel.commailto:yao.jin@intel.com

Thanks Jin Yao

From: Hicham GHORAYEB [mailto:notifications@github.com] Sent: Wednesday, May 18, 2016 3:42 PM To: 01org/numatop Cc: Jin, Yao; Comment Subject: Re: [01org/numatop] NumaTOP failed to start (#19)

With the root there is a core dumped:

ibm001srv root /nfshome/nonuser/autoengine/ bash$ numatop NumaTOP is starting ... Segmentation fault (core dumped)

— You are receiving this because you commented. Reply to this email directly or view it on GitHubhttps://github.com/01org/numatop/issues/19#issuecomment-219950202