search

openvstorage / framework

The Framework is a set of components and tools which brings the user an interface (GUI / API) to setup, extend and manage an Open vStorage platform.
Other
27 stars 23 forks source link

Zombie processes #753

Closed PeterNashaat closed 7 years ago

PeterNashaat commented 8 years ago

Every now and then we get an ovs module warning about one or more zombie processes ,sometimes it goes away and other times we have to kill the parent process to get rid of the zombie processes

wimpers commented 8 years ago

Hi Peter,

can you provide some more info? Which processes, can you add relevant logs?

PeterNashaat commented 8 years ago

Hi wim like now in cpu-05 be-scale-3 

root@cpu-05:~# ps aux | grep 'Z'
USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
ovs      11117  0.0  0.0      0     0 ?        Zs   14:30   0:00 [ssh] <defunct>
ovs      11118  0.0  0.0      0     0 ?        Zs   14:30   0:00 [ssh] <defunct>

and for the parent Processe 

root@cpu-05:~# ps aux | grep 17655
root     16442  0.0  0.0  11748  2220 pts/9    S+   14:50   0:00 grep --color=auto 17655
ovs      17655  0.0  0.2 760380 71600 ?        S    Jul31   0:42 python2.7 /usr/bin/celery worker -A ovs.celery_run -l info -Q ovs_generic,ovs_vJzYlMYEPAWj0xYN --autoscale=8,2 -Ofair
wimpers commented 8 years ago

@PeterNashaat do you know which action triggered it or is able to reproduce this? Is this Fargo or Eugene?

PeterNashaat commented 8 years ago

It keeps happening in all environments ,and it's on du-conv-3 right now ,and this is a Eugene env but we don't know how to reproduce it 

root@du-conv-3-01:~# ps aux | grep 'Z'
USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
ovs      16946  0.0  0.0      0     0 ?        Zs   10:30   0:00 [ssh] <defunct>
ovs      16951  0.0  0.0      0     0 ?        Zs   10:30   0:00 [ssh] <defunct>

and the Parent 

root@du-conv-3-01:~# cat /proc/16951/status | grep -i ppid
PPid:   16555
root@du-conv-3-01:~# ps aux | grep 16555
ovs      16555  0.1  0.0 689136 61352 ?        S    10:29   0:03 python2.7 /usr/bin/celery worker -A ovs.celery_run -l info -Q ovs_generic,ovs_Hp5mmtLSgOpmtHmm,ovs_masters --autoscale=8,2 -Ofair
wimpers commented 8 years ago

@pploegaert can you check if this also happens on our autotest env? Can we keep a eye on this on the OVH env?

khenderick commented 7 years ago

Put to verification since it might have been fixed by #820

JeffreyDevloo commented 7 years ago

Information

Found zombieprocesses on the Fargo environment that I installed for testing purposes right after the install (lingering SSH just as in described in the issue). Could not find the same zombie processes since #820 had been packaged.

Steps

Output

Found no zombie processes for the three installs that I did

Test result

Test passed. I have based my test result on the Fargo installation that would always leave a ssh zombie process.

Packages