metrics_monitor doesn't work in MSS 2017R2

[Mindaugas_R_]

Hello,

recently I have installed MSS 2017R2 on a new supermicro system X11SSV-M4F ( CPU: intel xeon E3-1585 v5, chipset: intel C236, OS CentOS 7.2.1511).

All works fine except metrics_monitor tool, which fails with this message: ERROR: Failed to subscribe for metrics, error code 1

And I can't find out what is wrong.

For example, I have another system running MSS 2017R2, it is intel NUC6i7KYK (mini pc) and here metrics_monitor runs fine. The graphics on both systems is the same Iris Pro 580.

Bellow you will find additional information:

[root@transcoder2 sample]# cat /etc/centos-release
CentOS Linux release 7.2.1511 (Core)

[root@transcoder2 sample]# uname -ar
Linux transcoder2 3.10.0-327.el7.x86_64 #1 SMP Thu Nov 19 22:10:57 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux

[root@transcoder2 sample]# lspci -nn -s 00:02.0
00:02.0 VGA compatible controller [0300]: Intel Corporation Device [8086:193a] (rev 09)

[root@transcoder2 sample]# lsmod | grep 'i915'
i915                 1325693  9
drm_ukmd_kms_helper   141060  1 i915
drm_ukmd              369649  3 i915,drm_ukmd_kms_helper
drm_ukmd_compat       109279  1 i915
i2c_algo_bit           13413  3 ast,igb,i915
video                  24400  1 i915
i2c_core               40582  11 ast,drm,igb,i915,i2c_i801,i2c_hid,ipmi_ssif,drm_kms_helper,i2c_algo_bit,drm_ukmd,drm_ukmd_kms_helper

[root@transcoder2 sample]# vainfo | grep -v unknown
error: can't connect to X server!
libva info: VA-API version 0.99.0
libva info: va_getDriverName() returns 0
libva info: User requested driver 'iHD'
libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so
libva info: Found init function __vaDriverInit_0_32
libva info: va_openDriver() returns 0
vainfo: VA-API version: 0.99 (libva 1.67.0.pre1)
vainfo: Driver version: 16.5.1.59511-ubit
vainfo: Supported profile and entrypoints
      VAProfileH264ConstrainedBaseline: VAEntrypointVLD
      VAProfileH264ConstrainedBaseline: VAEntrypointEncSlice
      VAProfileH264Main               : VAEntrypointVLD
      VAProfileH264Main               : VAEntrypointEncSlice
      VAProfileH264High               : VAEntrypointVLD
      VAProfileH264High               : VAEntrypointEncSlice
      VAProfileMPEG2Simple            : VAEntrypointEncSlice
      VAProfileMPEG2Simple            : VAEntrypointVLD
      VAProfileMPEG2Main              : VAEntrypointEncSlice
      VAProfileMPEG2Main              : VAEntrypointVLD
      VAProfileVC1Advanced            : VAEntrypointVLD
      VAProfileVC1Main                : VAEntrypointVLD
      VAProfileVC1Simple              : VAEntrypointVLD
      VAProfileJPEGBaseline           : VAEntrypointVLD
      VAProfileJPEGBaseline           : VAEntrypointEncPicture
      VAProfileVP8Version0_3          : VAEntrypointEncSlice
      VAProfileVP8Version0_3          : VAEntrypointVLD
      VAProfileHEVCMain               : VAEntrypointVLD
      VAProfileHEVCMain               : VAEntrypointEncSlice
      VAProfileNone                   : VAEntrypointVideoProc

[root@transcoder2 sample]# ./run.sh
ERROR: Failed to subscribe for metrics, error code 1

It will be nice to receive some suggestions from an expert

Regards,

Mindaugas

[rasko_N_]

Hello,

I have the same problem with the same proccessor (1585) but with E3-1275 no problems, works very well. Very strange, why anybody from Intel do not answer?

Regards,

Rashad

[rasko_N_]

Dear all,

Problem solved! Please try to update or downgrade gcc and g++ to version 4.9 and problem will solved.

Rashad

[Mindaugas_R_]

Hi Rashad,

glad to hear it. Could you give some more details, what have you changed step by step.

I have installed devtoolset-3 which includes gcc and g++ 4.9.2 and have recompiled metrics_monitor with that version but nothing have changed.

Regards,

Mindaugas

[rasko_N_]

hi Mindaugas,

i`m using Ubuntu 16.04 LTS  with gcc and g++ 4.9 

Rashad

[rasko_N_]

After upgraded gcc and g++ try to remove libdrm, libva and recompile again from /opt/intel/mediasdk/opensource/lib..

[rasko_N_]

After reboot same problem again. Very strange :(

[Mark_L_Intel1]

This is more like a software issue since the other people on the old platform got it.

https://software.intel.com/en-us/forums/intel-media-sdk/topic/540163#comment-1905526

Mark

[koby_s_]

Hi
I am having a similar problem on an Intel(R) Xeon(R) CPU E3-1285L v4 @ 3.40GHz machine, OS: CentOS 7.2.1511.
Hardware acceleration (decoding/vpp/encoding) is working well. Metrics query always fails to subscribe.

On the other hand, it works perfectly fine on a newer machine: Intel(R) Xeon(R) CPU E3-1275 v5 @ 3.60GHz, OS: CentOS 7.2.1511.

It can't be a GCC version problem (at least not on CentOS 7.2.1511) since it's working on that OS.

Did someone made a progress and managed to fix that issue?

Thank you

Koby

[Mark_L_Intel1]

Hi guys,

Any of you could try the following command to and post the log file?

[root@localhost _bin]# strace ./metrics_monitor

Mark

[benjamin_d_3]

Hello, here is strace ./metrics_monitor:

[root@gputest2 _bin]# strace ./metrics_monitor
execve("./metrics_monitor", ["./metrics_monitor"], 0x7ffe9a633c20 /* 31 vars */) = 0
brk(NULL)                               = 0x1db7000
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f7ea2718000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=84042, ...}) = 0
mmap(NULL, 84042, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7f7ea2703000
close(3)                                = 0
open("/lib64/tls/x86_64/libcttmetrics.so", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
stat("/lib64/tls/x86_64", 0x7ffc381c61f0) = -1 ENOENT (No such file or directory)
open("/lib64/tls/libcttmetrics.so", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
stat("/lib64/tls", {st_mode=S_IFDIR|0555, st_size=4096, ...}) = 0
open("/lib64/x86_64/libcttmetrics.so", O_RDONLY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
stat("/lib64/x86_64", 0x7ffc381c61f0)   = -1 ENOENT (No such file or directory)
open("/lib64/libcttmetrics.so", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\300\r\0\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=14624, ...}) = 0
mmap(NULL, 2112176, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f7ea22f1000
mprotect(0x7f7ea22f4000, 2093056, PROT_NONE) = 0
mmap(0x7f7ea24f3000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x2000) = 0x7f7ea24f3000
close(3)                                = 0
open("/lib64/libstdc++.so.6", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\2\1\1\3\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\200\270\10\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=1590720, ...}) = 0
mmap(NULL, 3699296, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f7ea1f69000
mprotect(0x7f7ea20e1000, 2097152, PROT_NONE) = 0
mmap(0x7f7ea22e1000, 49152, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x178000) = 0x7f7ea22e1000
mmap(0x7f7ea22ed000, 12896, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7f7ea22ed000
close(3)                                = 0
open("/lib64/libm.so.6", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\2\1\1\3\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0 W\0\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=1169408, ...}) = 0
mmap(NULL, 3178728, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f7ea1c60000
mprotect(0x7f7ea1d68000, 2093056, PROT_NONE) = 0
mmap(0x7f7ea1f67000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x107000) = 0x7f7ea1f67000
close(3)                                = 0
open("/lib64/libgcc_s.so.1", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\2\1\1\0\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\240*\0\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=92872, ...}) = 0
mmap(NULL, 2188336, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f7ea1a49000
mprotect(0x7f7ea1a5f000, 2093056, PROT_NONE) = 0
mmap(0x7f7ea1c5e000, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x15000) = 0x7f7ea1c5e000
close(3)                                = 0
open("/lib64/libc.so.6", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\2\1\1\3\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0 \5\2\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=2115824, ...}) = 0
mmap(NULL, 3955040, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7f7ea1683000
mprotect(0x7f7ea1840000, 2093056, PROT_NONE) = 0
mmap(0x7f7ea1a3f000, 24576, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x1bc000) = 0x7f7ea1a3f000
mmap(0x7f7ea1a45000, 14688, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7f7ea1a45000
close(3)                                = 0
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f7ea2701000
mmap(NULL, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f7ea26fe000
arch_prctl(ARCH_SET_FS, 0x7f7ea26fe740) = 0
mprotect(0x7f7ea1a3f000, 16384, PROT_READ) = 0
mprotect(0x7f7ea1c5e000, 4096, PROT_READ) = 0
mprotect(0x7f7ea1f67000, 4096, PROT_READ) = 0
mprotect(0x7f7ea22e1000, 40960, PROT_READ) = 0
mprotect(0x7f7ea24f3000, 4096, PROT_READ) = 0
mprotect(0x601000, 4096, PROT_READ)     = 0
mprotect(0x7f7ea271a000, 4096, PROT_READ) = 0
munmap(0x7f7ea2703000, 84042)           = 0
brk(NULL)                               = 0x1db7000
brk(0x1de9000)                          = 0x1de9000
rt_sigaction(SIGINT, {sa_handler=0x401020, sa_mask=[INT], sa_flags=SA_RESTORER|SA_RESTART, sa_restorer=0x7f7ea16b8990}, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=0}, 8) = 0
getuid()                                = 0
open("/sys/bus/pci/devices", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFDIR|0755, st_size=0, ...}) = 0
getdents(3, /* 23 entries */, 32768)    = 720
open("/sys/bus/pci/devices/0000:00:00.0/class", O_RDONLY) = 4
fstat(4, {st_mode=S_IFREG|0444, st_size=4096, ...}) = 0
read(4, "0x060000\n", 4096)             = 9
close(4)                                = 0
open("/sys/bus/pci/devices/0000:00:02.0/class", O_RDONLY) = 4
fstat(4, {st_mode=S_IFREG|0444, st_size=4096, ...}) = 0
read(4, "0x038000\n", 4096)             = 9
close(4)                                = 0
open("/sys/bus/pci/devices/0000:00:02.0/vendor", O_RDONLY) = 4
fstat(4, {st_mode=S_IFREG|0444, st_size=4096, ...}) = 0
read(4, "0x8086\n", 4096)               = 7
close(4)                                = 0
close(3)                                = 0
open("/sys/kernel/debug/dri/0/i915_ringstats", O_RDONLY) = -1 ENOENT (No such file or directory)
open("/sys/kernel/debug/dri/128/i915_ringstats", O_RDONLY) = 3
fstat(3, {st_mode=S_IFREG|0444, st_size=0, ...}) = 0
read(3, "render ring:0:-1:-1:-1:-1:-1\nbsd"..., 4096) = 152
read(3, "", 4096)                       = 0
close(3)                                = 0
open("/sys/class/drm/card0/gt_act_freq_mhz", O_RDONLY) = -1 ENOENT (No such file or directory)
write(2, "ERROR: Failed to subscribe for m"..., 53ERROR: Failed to subscribe for metrics, error code 1
) = 53
exit_group(1)                           = ?
+++ exited with 1 +++

[Mark_L_Intel1]

Thanks so much,

We have found this problem and the solution will be available in the next release.

The problem is: 

("/sys/class/drm/card0/gt_act_freq_mhz") is not exist on Intel(R) Xeon(R) E3-1285 v4 , so the configuration failed.

As a work around before the new release, you can apply the attached patch to cttmetrics_sample.cpp and recompile it.

Mark

[benjamin_d_3]

Thank you for the patch Mark.

Any idea when the MSS R3 will be released ? 

[Mindaugas_R_]

Thank you very much for the patch, Mark. At least I can see the GPU load now.

[Mark_L_Intel1]

Hi Benjamin,

I am glad to see the issue is fixed, Let me know if there is still a problem.

As far as I can see, it should be release quite soon, I can't tell you the exact date, but it should be within a month or two.

Mark