settingsLogin | Registersettings

[openstack-dev] Segmentation Fault in nova-compute service on ARM platform.

0 votes

Hi Guys,

I am porting openstack on arm platform, but while running nova-compute
service on it throws segmentaion fault.

Below are the logs:

2015-06-05 09:24:12.195 11374 INFO nova.virt.driver [-] Loading compute
driver 'libvirt.LibvirtDriver'
2015-06-05 09:24:13.483 11374 INFO nova.openstack.common.rpc.common
[req-ba16a115-ed64-4fb5-b13e-f28e03a28b24 None None] Connected to AMQP
server on 192.168.1.100:5672
2015-06-05 09:24:13.600 11374 INFO nova.openstack.common.rpc.common
[req-ba16a115-ed64-4fb5-b13e-f28e03a28b24 None None] Connected to AMQP
server on 192.168.1.100:5672
2015-06-05 09:24:13.808 11374 AUDIT nova.service [-] Starting compute node
(version 2013.2.2)

****Program received signal SIGSEGV, Segmentation fault.
0x76f244d4 in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0****

While tracing the same with gdb: I am getting corrupt stack:

""(gdb) bt

0 0x76f244d4 in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0

1 0x76f2ab78 in PyEval_EvalCodeEx () from /usr/lib/libpython2.7.so.1.0

2 0x76f2905c in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0

3 0x76f2ab78 in PyEval_EvalCodeEx () from /usr/lib/libpython2.7.so.1.0

4 0x76eaf690 in ?? () from /usr/lib/libpython2.7.so.1.0

Cannot access memory at address 0x0

5 0x76eaf690 in ?? () from /usr/lib/libpython2.7.so.1.0

Cannot access memory at address 0x0
Backtrace stopped: previous frame identical to this frame (corrupt stack?)""

The function PyEval_EvalFrameEx () is present in ceval.c

Also I am using havana release of openstack. I am new to openstack just
tried openstack nodes on x-86 platform.
Any help in order to solve the above issue is appreciated.

Regards
Pradeep Kumar


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
asked Jul 7, 2015 in openstack-dev by Pradeep_kumar (280 points)   1 2

7 Responses

0 votes

Havana is very very old and no longer supported. Why are you porting
that release in particular?

Michael

On Tue, Jul 7, 2015 at 3:51 PM, Pradeep kumar
topradeepyaduvanshi@gmail.com wrote:
Hi Guys,

I am porting openstack on arm platform, but while running nova-compute
service on it throws segmentaion fault.

Below are the logs:

2015-06-05 09:24:12.195 11374 INFO nova.virt.driver [-] Loading compute
driver 'libvirt.LibvirtDriver'
2015-06-05 09:24:13.483 11374 INFO nova.openstack.common.rpc.common
[req-ba16a115-ed64-4fb5-b13e-f28e03a28b24 None None] Connected to AMQP
server on 192.168.1.100:5672
2015-06-05 09:24:13.600 11374 INFO nova.openstack.common.rpc.common
[req-ba16a115-ed64-4fb5-b13e-f28e03a28b24 None None] Connected to AMQP
server on 192.168.1.100:5672
2015-06-05 09:24:13.808 11374 AUDIT nova.service [-] Starting compute node
(version 2013.2.2)

****Program received signal SIGSEGV, Segmentation fault.
0x76f244d4 in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0****

While tracing the same with gdb: I am getting corrupt stack:

""(gdb) bt

0 0x76f244d4 in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0

1 0x76f2ab78 in PyEval_EvalCodeEx () from /usr/lib/libpython2.7.so.1.0

2 0x76f2905c in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0

3 0x76f2ab78 in PyEval_EvalCodeEx () from /usr/lib/libpython2.7.so.1.0

4 0x76eaf690 in ?? () from /usr/lib/libpython2.7.so.1.0

Cannot access memory at address 0x0

5 0x76eaf690 in ?? () from /usr/lib/libpython2.7.so.1.0

Cannot access memory at address 0x0
Backtrace stopped: previous frame identical to this frame (corrupt stack?)""

The function PyEval_EvalFrameEx () is present in ceval.c

Also I am using havana release of openstack. I am new to openstack just
tried openstack nodes on x-86 platform.
Any help in order to solve the above issue is appreciated.

Regards
Pradeep Kumar


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

--
Rackspace Australia


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
responded Jul 7, 2015 by Michael_Still (16,180 points)   3 5 13
0 votes

Hi michael,

I did tried icehouse release but getting same error.
I think there is some issue with python code but how to resolve that need
some guidance on the same.
On Jul 7, 2015 11:32 AM, "Michael Still" mikal@stillhq.com wrote:

Havana is very very old and no longer supported. Why are you porting
that release in particular?

Michael

On Tue, Jul 7, 2015 at 3:51 PM, Pradeep kumar
topradeepyaduvanshi@gmail.com wrote:

Hi Guys,

I am porting openstack on arm platform, but while running nova-compute
service on it throws segmentaion fault.

Below are the logs:

2015-06-05 09:24:12.195 11374 INFO nova.virt.driver [-] Loading compute
driver 'libvirt.LibvirtDriver'
2015-06-05 09:24:13.483 11374 INFO nova.openstack.common.rpc.common
[req-ba16a115-ed64-4fb5-b13e-f28e03a28b24 None None] Connected to AMQP
server on 192.168.1.100:5672
2015-06-05 09:24:13.600 11374 INFO nova.openstack.common.rpc.common
[req-ba16a115-ed64-4fb5-b13e-f28e03a28b24 None None] Connected to AMQP
server on 192.168.1.100:5672
2015-06-05 09:24:13.808 11374 AUDIT nova.service [-] Starting compute
node
(version 2013.2.2)

****Program received signal SIGSEGV, Segmentation fault.
0x76f244d4 in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0****

While tracing the same with gdb: I am getting corrupt stack:

""(gdb) bt

0 0x76f244d4 in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0

1 0x76f2ab78 in PyEval_EvalCodeEx () from /usr/lib/libpython2.7.so.1.0

2 0x76f2905c in PyEval_EvalFrameEx () from /usr/lib/libpython2.7.so.1.0

3 0x76f2ab78 in PyEval_EvalCodeEx () from /usr/lib/libpython2.7.so.1.0

4 0x76eaf690 in ?? () from /usr/lib/libpython2.7.so.1.0

Cannot access memory at address 0x0

5 0x76eaf690 in ?? () from /usr/lib/libpython2.7.so.1.0

Cannot access memory at address 0x0
Backtrace stopped: previous frame identical to this frame (corrupt
stack?)""

The function PyEval_EvalFrameEx () is present in ceval.c

Also I am using havana release of openstack. I am new to openstack just
tried openstack nodes on x-86 platform.
Any help in order to solve the above issue is appreciated.

Regards
Pradeep Kumar


OpenStack Development Mailing List (not for usage questions)
Unsubscribe:
OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

--
Rackspace Australia


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
responded Jul 7, 2015 by Pradeep_kumar (280 points)   1 2
0 votes

On 07/07/2015 07:07 AM, Pradeep kumar wrote:
Hi michael,

I did tried icehouse release but getting same error.
I think there is some issue with python code but how to resolve that
need some guidance on the same.

It is extremely unlikely that anything in the OpenStack code is
responsible for a segfault in python. This seems like your python
interpreter being broken on your processor / os. I think you will need
to solve that much deeper problem.

-Sean

--
Sean Dague
http://dague.net


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
responded Jul 7, 2015 by Sean_Dague (66,200 points)   4 8 14
0 votes

Ohk sean
I am Using a platform with low configuration.IF u can help regarding some
achitectural changes actually i am stuck at same point from last 7 days.
Thanks sean
On Jul 7, 2015 5:11 PM, "Sean Dague" sean@dague.net wrote:

On 07/07/2015 07:07 AM, Pradeep kumar wrote:

Hi michael,

I did tried icehouse release but getting same error.
I think there is some issue with python code but how to resolve that
need some guidance on the same.

It is extremely unlikely that anything in the OpenStack code is
responsible for a segfault in python. This seems like your python
interpreter being broken on your processor / os. I think you will need
to solve that much deeper problem.

    -Sean

--
Sean Dague
http://dague.net


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
responded Jul 7, 2015 by Pradeep_kumar (280 points)   1 2
0 votes

I suggest you verify your platform bottom up. In this case make sure the virtualization layer works
before integrating openstack. Perhaps there is a test suite for libvirt or something that you could try.
/Hans

On 2015-07-07 14:55, Pradeep kumar wrote:
Ohk sean
I am Using a platform with low configuration.IF u can help regarding some achitectural changes
actually i am stuck at same point from last 7 days.
Thanks sean

On Jul 7, 2015 5:11 PM, "Sean Dague" <sean@dague.net sean@dague.net> wrote:

On 07/07/2015 07:07 AM, Pradeep kumar wrote:
 > Hi michael,
 >
 > I did tried icehouse release but getting same error.
 > I think there is some issue with python code but how to resolve that
 > need some guidance on the same.

It is extremely unlikely that anything in the OpenStack code is
responsible for a segfault in python. This seems like your python
interpreter being broken on your processor / os. I think you will need
to solve that much deeper problem.

         -Sean

--
Sean Dague
http://dague.net

__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe

http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
responded Jul 8, 2015 by Hans_Feldt (300 points)   1 1
0 votes

Hi Hans,
I did traced the function responsible for segmentation fault itz in file
ceval.c. find below

    if (*next_instr == SETUP_FINALLY) {
            /* Make the last opcode before
               a try: finally: block uninterruptible. */
            goto fast_next_opcode;
        }

After commenting above condition everything works fine, But need a
permanent workaround and some pointers
how to proceed further. Till now not able to understand the code.
Thanks guys.

Regards
Pradeep Kumar

On Wed, Jul 8, 2015 at 3:12 PM, Hans Feldt hans.feldt@ericsson.com wrote:

I suggest you verify your platform bottom up. In this case make sure the
virtualization layer works before integrating openstack. Perhaps there is a
test suite for libvirt or something that you could try.
/Hans

On 2015-07-07 14:55, Pradeep kumar wrote:

Ohk sean
I am Using a platform with low configuration.IF u can help regarding some
achitectural changes
actually i am stuck at same point from last 7 days.
Thanks sean

On Jul 7, 2015 5:11 PM, "Sean Dague" <sean@dague.net > wrote:

On 07/07/2015 07:07 AM, Pradeep kumar wrote:
 > Hi michael,
 >
 > I did tried icehouse release but getting same error.
 > I think there is some issue with python code but how to resolve

that

need some guidance on the same.

It is extremely unlikely that anything in the OpenStack code is
responsible for a segfault in python. This seems like your python
interpreter being broken on your processor / os. I think you will need
to solve that much deeper problem.

         -Sean

--
Sean Dague
http://dague.net


OpenStack Development Mailing List (not for usage questions)
Unsubscribe:
OpenStack-dev-request@lists.openstack.org?subject:unsubscribe


OpenStack Development Mailing List (not for usage questions)
Unsubscribe:
OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
responded Jul 9, 2015 by Pradeep_kumar (280 points)   1 2
0 votes

On Thu, Jul 09, 2015 at 05:19:06PM +0530, Pradeep kumar wrote:
Hi Hans,
I did traced the function responsible for segmentation fault itz in file
ceval.c. find below

    if (*next_instr == SETUP_FINALLY) {
            /* Make the last opcode before
               a try: finally: block uninterruptible. */
            goto fast_next_opcode;
        }

After commenting above condition everything works fine, But need a
permanent workaround and some pointers
how to proceed further. Till now not able to understand the code.
Thanks guys.

We didn't see this exact problem BUT when runing OpenStack on a new platform we
hit problems in qemu, libvirt, greenthreads, eventlet, erlang and probably a
few others.

I think you need to (as Han's suggested) verify each step from "system booting" to OpenStack.

Look at each of the required components and verify that ther tests suits pass
for example does eventlet work on your system?

If a test fails then understand why. You may not need to fix that specific
case if it's of no use to OpenStack "right now".

Good luck.

Yours Tony.


OpenStack Development Mailing List (not for usage questions)
Unsubscribe: OpenStack-dev-request@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

responded Jul 9, 2015 by Tony_Breeds (19,660 points)   3 6 9
...