6:45 a.m.
On 12/19/2013 02:03 AM, Guang Ya GY Liu wrote:
Hi,
I was following this install guide to install OpenStack with RDO:
http://openstack.redhat.com/Quickstart
Hi,
I'm going to move this thread over to rdo-list, since that's where the
RDO community members hang out.
After install and OpenStack start, every time I boot a new instance,
there will be a new nova-compute start up, after I create two instance,
there are three nova-compute, do you know why the instance create caused
new nova-compute start?
[root@db03b04 network-scripts(keystone_admin)]# ps -ef | grep nova-com
nova 5174 6099 0 01:07 ? 00:00:00 /usr/bin/python
/usr/bin/nova-compute --logfile /var/log/nova/compute.log
<<<<<<<<<< Why?
nova 6099 1 0 Dec18 ? 00:07:47 /usr/bin/python
/usr/bin/nova-compute --logfile /var/log/nova/compute.log
nova 6553 6099 0 Dec18 ? 00:00:00 /usr/bin/python
/usr/bin/nova-compute --logfile /var/log/nova/compute.log
<<<<<<<<<< Why?
Russell, can you provide a quick explanation here?
Cheers,
Perry
12:43 p.m.
New subject: [Rdo-list] [rhos-list] RDO problem
On 12/19/2013 07:45 AM, Perry Myers wrote:
On 12/19/2013 02:03 AM, Guang Ya GY Liu wrote:
> Hi,
>
> I was following this install guide to install OpenStack with RDO:
> http://openstack.redhat.com/Quickstart
Hi,
I'm going to move this thread over to rdo-list, since that's where the
RDO community members hang out.
> After install and OpenStack start, every time I boot a new instance,
> there will be a new nova-compute start up, after I create two instance,
> there are three nova-compute, do you know why the instance create caused
> new nova-compute start?
>
> [root@db03b04 network-scripts(keystone_admin)]# ps -ef | grep nova-com
> nova 5174 6099 0 01:07 ? 00:00:00 /usr/bin/python
> /usr/bin/nova-compute --logfile /var/log/nova/compute.log
<<<<<<<<<< Why?
> nova 6099 1 0 Dec18 ? 00:07:47 /usr/bin/python
> /usr/bin/nova-compute --logfile /var/log/nova/compute.log
> nova 6553 6099 0 Dec18 ? 00:00:00 /usr/bin/python
> /usr/bin/nova-compute --logfile /var/log/nova/compute.log
<<<<<<<<<< Why?
Russell, can you provide a quick explanation here?
nova-compute spawns child processes, but not to run multiple instances
of nova-compute itself.
Perhaps it's just something inaccurate with the process list where the
command part didn't get updated... not sure. In this case 6099 is the
normal parent nova-compute. The others are suspect.
Can you grab an strace log from them? Something like ...
# strace -p 5174 -o 5174.txt
--
Russell Bryant
12:46 p.m.
New subject: [Rdo-list] [rhos-list] RDO problem
nova-compute spawns child processes, but not to run multiple
instances
of nova-compute itself.
Perhaps it's just something inaccurate with the process list where the
command part didn't get updated... not sure. In this case 6099 is the
normal parent nova-compute. The others are suspect.
Can you grab an strace log from them? Something like ...
# strace -p 5174 -o 5174.txt
Also do this, which might be more obvious if it's what we think it is:
ls -l /proc/5174/exe
--Dan
4:23 p.m.
New subject: [Rdo-list] [rhos-list] RDO problem
Thanks Russell and Dan, the following are the output that you need.
[root@db03b04 ~]# cat 5174.txt
restart_syscall(<... resuming interrupted call ...>) = 0
kill(5173, SIG_0) = 0
kill(6099, SIG_0) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
rt_sigaction(SIGCHLD, NULL, {SIG_DFL, [], SA_RESTORER, 0x7f9827095500}, 8)
= 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
nanosleep({2, 0}, 0x7f97ebffcb80) = 0
kill(5173, SIG_0) = 0
kill(6099, SIG_0) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
rt_sigaction(SIGCHLD, NULL, {SIG_DFL, [], SA_RESTORER, 0x7f9827095500}, 8)
= 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
nanosleep({2, 0}, 0x7f97ebffcb80) = 0
kill(5173, SIG_0) = 0
kill(6099, SIG_0) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
rt_sigaction(SIGCHLD, NULL, {SIG_DFL, [], SA_RESTORER, 0x7f9827095500}, 8)
= 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
nanosleep({2, 0}, 0x7f97ebffcb80) = 0
kill(5173, SIG_0) = 0
kill(6099, SIG_0) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
rt_sigaction(SIGCHLD, NULL, {SIG_DFL, [], SA_RESTORER, 0x7f9827095500}, 8)
= 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
nanosleep({2, 0}, 0x7f97ebffcb80) = 0
kill(5173, SIG_0) = 0
kill(6099, SIG_0) = 0
rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
[root@db03b04 ~]# ls -l /proc/5174/exe
lrwxrwxrwx 1 nova nova 0 Dec 19 01:27 /proc/5174/exe -> /usr/bin/python
Thanks,
Guangya
From:
Dan Smith <dansmith(a)redhat.com>
To:
Russell Bryant <rbryant(a)redhat.com>, Perry Myers <pmyers(a)redhat.com>,
Guang Ya GY Liu/China/IBM@IBMCN, rdo-list <rdo-list(a)redhat.com>
Date:
2013/12/20 02:46
Subject:
Re: [rhos-list] RDO problem
nova-compute spawns child processes, but not to run multiple
instances
of nova-compute itself.
Perhaps it's just something inaccurate with the process list where the
command part didn't get updated... not sure. In this case 6099 is the
normal parent nova-compute. The others are suspect.
Can you grab an strace log from them? Something like ...
# strace -p 5174 -o 5174.txt
Also do this, which might be more obvious if it's what we think it is:
ls -l /proc/5174/exe
--Dan
4338
days inactive
4338
days old
3 comments
4 participants
participants (4)
-
Dan Smith -
Guang Ya GY Liu -
Perry Myers -
Russell Bryant