New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 725589 link

Starred by 1 user

Issue metadata

Status: Archived
Owner:
Closed: Jun 2017
Cc:
EstimatedDays: ----
NextAction: ----
OS: Chrome
Pri: ----
Type: Bug



Sign in to add a comment

guado moblab: frequent reboots due to low memory

Project Member Reported by lgoo...@chromium.org, May 23 2017

Issue description

Guado moblab 9202.70.0 testimage with 4GB RAM reboots multiple times per day when in use.

Log shows oom-killer active prior to reboot.
No evidence of security scan at the time.

With no test load on the moblab, free -m shows:

             total       used       free     shared    buffers     cached
Mem:          3883       3574        309         38          1        113
-/+ buffers/cache:       3460        423
Swap:         5689       2410       3278

Looking at top users of virtual memory, no tests running:

top -n 1 -o VIRT -c

Tasks: 222 total,   1 running, 221 sleeping,   0 stopped,   0 zombie
%Cpu(s):  6.6 us,  2.0 sy,  0.3 ni, 91.0 id,  0.2 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:   3977204 total,  3657972 used,   319232 free,     4036 buffers
KiB Swap:  5825980 total,  2467884 used,  3358096 free.   118948 cached Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND                                                                                
 3878 moblab    20   0 3374020   6768   2356 S   0.0  0.2   1:08.16 /usr/bin/python2 /usr/lib/devserver/devserver.py --production --static_dir=/mnt/mobla+ 
 1047 mysql     20   0 1421372  38696   4588 S   0.0  1.0   7:01.27 /usr/sbin/mysqld                                                                       
  541 root      20   0 1104584  49312   2420 S   0.0  1.2  22:55.41 python2 /usr/bin/mobmonitor                                                            
 1086 chronos   20   0  746660  33020   6076 S   0.0  0.8   8:00.55 /opt/google/chrome/chrome --type=renderer --enable-logging --log-level=1 --use-gl=egl+ 
  920 chronos   20   0  737360  24340   2980 S   0.0  0.6   0:02.22 /opt/google/chrome/chrome --type=renderer --enable-logging --log-level=1 --use-gl=egl+ 
  807 chronos   12  -8  547716  27968   9592 S   0.0  0.7   5:42.59 /opt/google/chrome/chrome --ui-prioritize-in-gpu-process --use-gl=egl --gpu-sandbox-f+ 
11175 apache    20   0  353376 130848   5916 S   0.0  3.3   0:08.72 apache2                                                                                
31779 apache    20   0  352096 134420   7784 S   0.0  3.4   0:03.17 apache2                                                                                
  335 apache    20   0  351840  29880   4984 S   0.0  0.8   0:07.71 apache2                                                                                
 3296 apache    20   0  351840  47912   6660 S   0.0  1.2   0:04.05 apache2                                                                                
 7410 apache    20   0  351840  23036   4648 S   0.0  0.6   0:07.37 apache2                                                                                
12599 apache    20   0  351840 118704   6676 S   0.0  3.0   0:03.61 apache2                                                                                
14363 apache    20   0  351840  23088   4636 S   0.0  0.6   0:08.03 apache2  

Note that devserver and mobmonitor virt usage is high.

The syslog from just prior to the oom-related crash is attached.


 
oom-crash.log
210 KB View Download
Upgrading the moblab from 4GB to 6GB seems to have completely eliminated this problem.

Comment 2 by krk@google.com, Jun 8 2017

Owner: haddowk@chromium.org
Status: Fixed (was: Untriaged)
Should be fixed in 59
https://chromium-review.googlesource.com/#/c/501989/
Cc: kohy@chromium.org

Comment 5 by dchan@chromium.org, Jan 22 2018

Status: Archived (was: Fixed)

Sign in to add a comment