HIGH CPU Fog Services after update r5029 v6759

[Tom Elliott]

Why this might affect the CPU? It looks to the DB to get timeout values (and logs). The DB is fully operational, but it’s not initiated by fog meaning the DB is not available. It returns an int of 0 for the sleep time. Put that into an infinite loop, (there’s at least 2 that manage start/restart of the services. It is told to sleep for some period of time already within the looping I’m referring to. However it’s set to 0 (meaning 0 seconds), so it get’s stuck in an infinite loop without ever actually starting.

[baggar11]

@Tom-Elliott said:

@baggar11, @Raymond-Bell I believe I may have fixed this now.
With any luck, this issue will be gone now. Please update and let me know.

Just updated to 6795, removed my “restart” lines from rc.local and rebooted. CPU was pegged again. Restarting the services manually brought cpu back down to idle.

[Tom Elliott]

@baggar11 as I understand it, this is still potentially due to starting the services before its dependencies have started. This is different from the pegged on update, I think.

Can you simply try reinstall and see if it’s still happening? Or did update always work, only reboot caused these issues?

[baggar11]

@Tom-Elliott said:

@baggar11 as I understand it, this is still potentially due to starting the services before its dependencies have started. This is different from the pegged on update, I think.

Can you simply try reinstall and see if it’s still happening? Or did update always work, only reboot caused these issues?

For my upgrades, anything after 6753 would peg my cpu after a reboot. During and after the upgrade process using the install.sh, everything was fine. It’s only after the reboot that the services would max out the cpu. Probably the infinite loop issue as you wrote about below.

[Tom Elliott]

@baggar11 Okay, I’ve added a sleep time of 10 seconds, just in case of these situations where the int returned is 0. We must have at least 1 second sleeptime ( I believe ), and this just isn’t happening.

The only word of caution I can think of, now, is while this might help with CPU Cycles, it will not make the FOG Services actually work properly as from Ubuntu’s standpoint the service is already running. While I could, potentially, come up with a way to restart the services more appropriately, I don’t know where to start at the moment.

[baggar11]

@Tom-Elliott said:

@baggar11 Okay, I’ve added a sleep time of 10 seconds, just in case of these situations where the int returned is 0. We must have at least 1 second sleeptime ( I believe ), and this just isn’t happening.

The only word of caution I can think of, now, is while this might help with CPU Cycles, it will not make the FOG Services actually work properly as from Ubuntu’s standpoint the service is already running. While I could, potentially, come up with a way to restart the services more appropriately, I don’t know where to start at the moment.

I’m not a coder so take this with a grain of salt. The following is a similar approach to another open source project I use.

Make a script to run as a cron job every 5 minutes. On 1st boot, the script essentially checks for things like MySQL and network being up, and if all checks out starts the FOG services. All other times it runs, it would really only check for FOG services running and then exit. heh…

Of course, it’s probably a waste of time as I assume systemd will probably take care of these things more intelligently than regular init and/or upstart. Which Fedora is already using and Ubuntu’s next LTS will be running.

I really appreciate you taking a look into these things Tom. FOG is a very cool project.

[boeleke]

@baggar11 & @Tom-Elliott

Like Baggar is saying. Thanks for taking your time to looking at these things.
Its just anoying that it is taking 100% CPU & not working properly.
Take your time but its probably a problem for all the ubuntu 14.04 - users after one update.

[Raymond Bell]

@Tom-Elliott After update to r5054 this morning
Looks like it is fixed now but i have not rebooted any of them…
Server

top - 07:57:51 up 23:03,  3 users,  load average: 134.00, 122.87, 122.12
Tasks: 351 total,  32 running, 312 sleeping,   0 stopped,   7 zombie
%Cpu(s): 55.3 us, 42.0 sy,  0.0 ni,  0.0 id,  0.0 wa,  0.0 hi,  2.7 si,  0.0 st
KiB Mem:   4355692 total,  3461944 used,   893748 free,    69372 buffers
KiB Swap:  1046524 total,    16896 used,  1029628 free.  2231712 cached Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
17801 mysql     20   0  386360  85396   9744 S  19.5  2.0   3:07.93 mysqld
20119 www-data  20   0  111272  19556  13116 R  11.3  0.4   0:04.27 apache2
21537 www-data  20   0       0      0      0 Z   9.9  0.0   0:01.06 apache2
21239 www-data  20   0  111188  19952  13344 R   8.6  0.5   0:02.65 apache2
19002 www-data  20   0  111044  19584  13116 R   7.0  0.4   0:08.43 apache2
18759 www-data  20   0  111296  19812  13092 R   6.6  0.5   0:13.16 apache2
19062 www-data  20   0  110988  19788  13752 R   6.6  0.5   0:07.04 apache2
20017 www-data  20   0  110996  18988  12572 S   6.6  0.4   0:04.30 apache2
18845 www-data  20   0  111300  19240  12688 R   6.3  0.4   0:13.16 apache2
18875 www-data  20   0  111256  20412  13732 R   6.3  0.5   0:12.60 apache2

Storage Node Master

top - 07:58:14 up 19:25,  2 users,  load average: 0.01, 0.06, 0.12
Tasks: 201 total,   1 running, 195 sleeping,   0 stopped,   5 zombie
%Cpu(s):  0.2 us,  0.2 sy,  0.0 ni, 99.7 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:   1014980 total,   975276 used,    39704 free,   106172 buffers
KiB Swap:  1037308 total,    22532 used,  1014776 free.   443116 cached Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
    1 root      20   0    4616   3552   2432 S   0.0  0.3   0:02.85 init
    2 root      20   0       0      0      0 S   0.0  0.0   0:00.00 kthreadd
    3 root      20   0       0      0      0 S   0.0  0.0   0:03.88 ksoftirqd/0
    5 root       0 -20       0      0      0 S   0.0  0.0   0:00.00 kworker/0:0H
    7 root      20   0       0      0      0 S   0.0  0.0   0:26.13 rcu_sched
    8 root      20   0       0      0      0 S   0.0  0.0   0:00.00 rcu_bh
    9 root      rt   0       0      0      0 S   0.0  0.0   0:00.02 migration/0
   10 root      rt   0       0      0      0 S   0.0  0.0   0:00.18 watchdog/0
   11 root      rt   0       0      0      0 S   0.0  0.0   0:00.20 watchdog/1
   12 root      rt   0       0      0      0 S   0.0  0.0   0:00.02 migration/1

Storage Node 2nd

top - 07:58:31 up 22:59,  3 users,  load average: 0.01, 0.02, 0.05
Tasks: 227 total,   1 running, 225 sleeping,   0 stopped,   1 zombie
%Cpu(s):  0.1 us,  0.0 sy,  0.0 ni, 99.9 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st
KiB Mem:   7735460 total,  7466792 used,   268668 free,    69996 buffers
KiB Swap:  7828476 total,        0 used,  7828476 free.  6881904 cached Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
    7 root      20   0       0      0      0 S   0.3  0.0   0:29.70 rcu_sched
11185 fog       20   0    5572   2852   2352 R   0.3  0.0   0:02.77 top
    1 root      20   0    4732   3844   2580 S   0.0  0.0   0:01.60 init
    2 root      20   0       0      0      0 S   0.0  0.0   0:00.00 kthreadd
    3 root      20   0       0      0      0 S   0.0  0.0   0:04.09 ksoftirqd/0
    5 root       0 -20       0      0      0 S   0.0  0.0   0:00.00 kworker/0:0H
    8 root      20   0       0      0      0 S   0.0  0.0   0:00.00 rcu_bh
    9 root      rt   0       0      0      0 S   0.0  0.0   0:00.01 migration/0
   10 root      rt   0       0      0      0 S   0.0  0.0   0:00.30 watchdog/0
   11 root      rt   0       0      0      0 S   0.0  0.0   0:00.32 watchdog/1
   12 root      rt   0       0      0      0 S   0.0  0.0   0:00.01 migration/1

[boeleke]

@Raymond-Bell

you are right. CPU is fixed indead … Till you boot

[Raymond Bell]

@boeleke said:

@Raymond-Bell

you are right. CPU is fixed indead … Till you boot

Ok i have not rebooted yet

[Tom Elliott]

@boeleke What?

You updated and still seeing CPU load issues? See the service_lib script now implicitly defines a sleep time if one cannot be found otherwise.

[boeleke]

@Tom-Elliott
thought i had the last version.
I just updated & rebooted. Now working fine for a couple of minutes.

think its ok now.

[baggar11]

Wanted to post another update. Updated to 6893 this morning. Removed my sleep and service stop commands from rc.local and rebooted. CPU seems to be in check again. Thanks for all your hard work!