HIGH CPU Fog Services after update r5029 v6759

Raymond Bell

@Raymond-Bell The rc.local stuff is also on the Storage Nodes?

yes i edited rc.local on Storage Nodes also

Tom Elliott

@Raymond-Bell Did you restart the services after adding the rc.local? I’m sorry if I’m asking obvious questions, but I really just need to make sure. Current update should help alleviate CPU usage as well though.

Raymond Bell

@Tom-Elliott said:

@Raymond-Bell Did you restart the services after adding the rc.local? I’m sorry if I’m asking obvious questions, but I really just need to make sure. Current update should help alleviate CPU usage as well though.

yes restarted and rebooted and also up to r5038 6777
will do it again to make sure

Raymond Bell

@Tom-Elliott Yes same result after

sudo service FOGMulticastManager stop
sudo service FOGImageReplicator stop
sudo service FOGSnapinReplicator stop
sudo service FOGScheduler stop
sudo service FOGPingHosts stop


sudo update-rc.d FOGMulticastManager disable
sudo update-rc.d FOGImageReplicator disable
sudo update-rc.d FOGSnapinReplicator disable
sudo update-rc.d FOGScheduler disable
sudo update-rc.d FOGPingHosts disable

sudo service FOGMulticastManager start
sudo service FOGImageReplicator start
sudo service FOGSnapinReplicator start
sudo service FOGScheduler start
sudo service FOGPingHosts start

top - 10:48:18 up  1:48,  2 users,  load average: 4.14, 4.92, 5.02
Tasks: 202 total,   6 running, 196 sleeping,   0 stopped,   0 zombie
%Cpu(s): 66.6 us, 32.7 sy,  0.0 ni,  0.0 id,  0.0 wa,  0.0 hi,  0.7 si,  0.0 st
KiB Mem:   1014980 total,   966204 used,    48776 free,    19852 buffers
KiB Swap:  1037308 total,    12316 used,  1024992 free.   529972 cached Mem

  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 5970 root      20   0   41920  20164  15184 R  45.6  2.0   0:07.37 FOGSnapinReplic
 6026 root      20   0   41920  20392  15408 R  41.3  2.0   0:05.85 FOGPingHosts
 5990 root      20   0   41920  20240  15256 R  39.6  2.0   0:06.34 FOGTaskSchedule
 5950 root      20   0   41920  20488  15504 R  39.3  2.0   0:06.90 FOGImageReplica
 5930 root      20   0   41920  20116  15132 R  34.3  2.0   0:05.32 FOGMulticastMan
 5779 fog       20   0   11284   3900   3128 S   0.3  0.4   0:00.04 sshd
    1 root      20   0    4732   3788   2548 S   0.0  0.4   0:02.83 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   0:00.78 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   0:05.27 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   0:00.01 migration/0
   10 root      rt   0       0      0      0 S   0.0  0.0   0:00.01 watchdog/0
   11 root      rt   0       0      0      0 S   0.0  0.0   0:00.01 watchdog/1
   12 root      rt   0       0      0      0 S   0.0  0.0   0:00.00 migration/1
   13 root      20   0       0      0      0 S   0.0  0.0   0:00.12 ksoftirqd/1

boeleke

alt text

Something weird. After the new update , my CPU performance from my server was going to 10-20%. After a reboot , he started again @100%

Pretty anoying

baggar11

I just realized after going through another git pull upgrade that the CPU spike will probably happen every time since the services are reinstalled. Here’s how I fixed my problem again, hopefully in a cleaner way. After upgrading to 6791, it seems to be working so far with a couple reboot tests.

Add this to your rc.local. Adjust sleep time based on your setup. My FOG setup is virtual, so 5 seconds seems to work well.

sleep 5
service FOGImageReplicator restart
service FOGMulticastManager restart
service FOGPingHosts restart
service FOGScheduler restart
service FOGSnapinReplicator restart

Tom Elliott

@baggar11, @Raymond-Bell I believe I may have fixed this now.

My issue – in theory?

I am moving almost EVERYTHING to static form (where I can). This allows me to get items back without having to initiate a whole class object. The problem, the DB has to be available for FOG to read its info. The services are checking for the item before DB is established. This was totally an over site and for that I’m sorry to all.

With any luck, this issue will be gone now. Please update and let me know.

Thanks.

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!

HIGH CPU Fog Services after update r5029 v6759

219

11.7k

17.1k

154.7k