Robin Lee Powell
2012-Jul-02 18:42 UTC
[Puppet Users] puppetmasterd continuously consuming high CPU, with many interrupts
So, I have a server at home that has four VMs running inside it.
All are managed via puppet. The physical host runs puppetmasterd.
I don''t recall noticing this before, but puppetmasterd has decided
to be kind of crazy. Here''s the physical host with no puppetmasterd
running:
top - 11:36:15 up 271 days, 15:16, 1 user, load average: 5.68, 5.50, 6.45
Tasks: 129 total, 1 running, 128 sleeping, 0 stopped, 0 zombie
Cpu(s): 3.6%us, 1.8%sy, 0.0%ni, 80.4%id, 14.3%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 8128776k total, 6991020k used, 1137756k free, 408756k buffers
Swap: 8388604k total, 552356k used, 7836248k free, 185220k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
10296 qemu 20 0 2366m 1.5g 8884 S 6.2 19.6 6:37.65 qemu-kvm
17334 qemu 20 0 2788m 1.7g 544 S 2.7 22.3 4576:25 qemu-kvm
9904 qemu 20 0 2358m 581m 8820 S 0.9 7.3 3:55.78 qemu-kvm
1 root 20 0 46880 8076 1372 S 0.0 0.1 0:27.00 systemd
2 root 20 0 0 0 0 S 0.0 0.0 0:10.48 kthreadd
3 root 20 0 0 0 0 S 0.0 0.0 322:04.84 ksoftirqd/0
6 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/0
7 root RT 0 0 0 0 S 0.0 0.0 0:11.57 watchdog/0
8 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/1
10 root 20 0 0 0 0 S 0.0 0.0 551:03.31 ksoftirqd/1
And here it is with puppetmasterd running:
top - 11:25:07 up 271 days, 15:05, 1 user, load average: 12.59, 8.68, 7.05
Tasks: 131 total, 3 running, 128 sleeping, 0 stopped, 0 zombie
Cpu(s): 15.2%us, 36.4%sy, 0.0%ni, 6.6%id, 39.7%wa, 0.0%hi, 2.0%si, 0.0%st
Mem: 8128776k total, 6830276k used, 1298500k free, 381356k buffers
Swap: 8388604k total, 555328k used, 7833276k free, 180096k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
10660 puppet 20 0 214m 107m 4040 S 61.9 1.3 8:46.81 puppetmasterd
3 root 20 0 0 0 0 S 21.4 0.0 320:38.54 ksoftirqd/0
10 root 20 0 0 0 0 R 20.2 0.0 549:30.88 ksoftirqd/1
10296 qemu 20 0 2470m 1.4g 8888 S 13.1 18.1 4:23.70 qemu-kvm
17334 qemu 20 0 2788m 1.7g 540 S 8.3 22.0 4574:54 qemu-kvm
9904 qemu 20 0 2422m 572m 8820 S 3.6 7.2 3:07.15 qemu-kvm
24980 qemu 20 0 1824m 1.4g 612 S 3.6 18.3 15046:11 qemu-kvm
12209 rlpowell 20 0 15256 1228 908 R 1.2 0.0 0:00.04 top
1 root 20 0 46880 7992 1356 S 0.0 0.1 0:26.97 systemd
2 root 20 0 0 0 0 S 0.0 0.0 0:10.48 kthreadd
The high CPU use by puppetmasterd is bad enough, but what makes me
be all like "wait, what?" is the ksoftirqd usage.
Puppet master version is 2.16.
This is *without* a client running; there''s no traffic on 8140
according to tcpdump, and there''s nothing happening in the log.
http://users.digitalkingdom.org/~rlpowell/media/public/puppetmasterd_strace.txt
has strace output; it''s pretty boring, but there are a few select
and rt_sigprocmask calls near the bottom.
I''m totally stumped here. Any ideas?
-Robin
--
http://singinst.org/ : Our last, best hope for a fantastic future.
.i ko na cpedu lo nu stidi vau loi jbopre .i danfu lu na go''i
li''u .e
lu go''i li''u .i ji''a go''i lu na''e
go''i li''u .e lu go''i na''i li''u .e
lu no''e go''i li''u .e lu to''e go''i
li''u .e lu lo mamta be do cu sofybakni li''u
--
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To post to this group, send email to puppet-users@googlegroups.com.
To unsubscribe from this group, send email to
puppet-users+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/puppet-users?hl=en.
Ashley Penney
2012-Jul-02 19:01 UTC
Re: [Puppet Users] puppetmasterd continuously consuming high CPU, with many interrupts
It might be totally unrelated but check for ksoftirqd and see if it''s running with high CPU. The leap second the other day caused all my puppetmasters to spike up to 100% CPU and other people had similar problems. I notice your server has 271 days uptime so it might not be the cause but it''s worth trying to either set the date with date -s or reboot the machine to see if it clears it up. On Mon, Jul 2, 2012 at 2:42 PM, Robin Lee Powell < rlpowell@digitalkingdom.org> wrote:> So, I have a server at home that has four VMs running inside it. > All are managed via puppet. The physical host runs puppetmasterd. > > I don''t recall noticing this before, but puppetmasterd has decided > to be kind of crazy. Here''s the physical host with no puppetmasterd > running: > > top - 11:36:15 up 271 days, 15:16, 1 user, load average: 5.68, 5.50, 6.45 > Tasks: 129 total, 1 running, 128 sleeping, 0 stopped, 0 zombie > Cpu(s): 3.6%us, 1.8%sy, 0.0%ni, 80.4%id, 14.3%wa, 0.0%hi, 0.0%si, > 0.0%st > Mem: 8128776k total, 6991020k used, 1137756k free, 408756k buffers > Swap: 8388604k total, 552356k used, 7836248k free, 185220k cached > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > 10296 qemu 20 0 2366m 1.5g 8884 S 6.2 19.6 6:37.65 qemu-kvm > 17334 qemu 20 0 2788m 1.7g 544 S 2.7 22.3 4576:25 qemu-kvm > 9904 qemu 20 0 2358m 581m 8820 S 0.9 7.3 3:55.78 qemu-kvm > 1 root 20 0 46880 8076 1372 S 0.0 0.1 0:27.00 systemd > 2 root 20 0 0 0 0 S 0.0 0.0 0:10.48 kthreadd > 3 root 20 0 0 0 0 S 0.0 0.0 322:04.84 ksoftirqd/0 > 6 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/0 > 7 root RT 0 0 0 0 S 0.0 0.0 0:11.57 watchdog/0 > 8 root RT 0 0 0 0 S 0.0 0.0 0:00.00 migration/1 > 10 root 20 0 0 0 0 S 0.0 0.0 551:03.31 ksoftirqd/1 > > And here it is with puppetmasterd running: > > top - 11:25:07 up 271 days, 15:05, 1 user, load average: 12.59, 8.68, > 7.05 > Tasks: 131 total, 3 running, 128 sleeping, 0 stopped, 0 zombie > Cpu(s): 15.2%us, 36.4%sy, 0.0%ni, 6.6%id, 39.7%wa, 0.0%hi, 2.0%si, > 0.0%st > Mem: 8128776k total, 6830276k used, 1298500k free, 381356k buffers > Swap: 8388604k total, 555328k used, 7833276k free, 180096k cached > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > 10660 puppet 20 0 214m 107m 4040 S 61.9 1.3 8:46.81 puppetmasterd > 3 root 20 0 0 0 0 S 21.4 0.0 320:38.54 ksoftirqd/0 > 10 root 20 0 0 0 0 R 20.2 0.0 549:30.88 ksoftirqd/1 > 10296 qemu 20 0 2470m 1.4g 8888 S 13.1 18.1 4:23.70 qemu-kvm > 17334 qemu 20 0 2788m 1.7g 540 S 8.3 22.0 4574:54 qemu-kvm > 9904 qemu 20 0 2422m 572m 8820 S 3.6 7.2 3:07.15 qemu-kvm > 24980 qemu 20 0 1824m 1.4g 612 S 3.6 18.3 15046:11 qemu-kvm > 12209 rlpowell 20 0 15256 1228 908 R 1.2 0.0 0:00.04 top > 1 root 20 0 46880 7992 1356 S 0.0 0.1 0:26.97 systemd > 2 root 20 0 0 0 0 S 0.0 0.0 0:10.48 kthreadd > > The high CPU use by puppetmasterd is bad enough, but what makes me > be all like "wait, what?" is the ksoftirqd usage. > > Puppet master version is 2.16. > > This is *without* a client running; there''s no traffic on 8140 > according to tcpdump, and there''s nothing happening in the log. > > > http://users.digitalkingdom.org/~rlpowell/media/public/puppetmasterd_strace.txt > has strace output; it''s pretty boring, but there are a few select > and rt_sigprocmask calls near the bottom. > > I''m totally stumped here. Any ideas? > > -Robin > > -- > http://singinst.org/ : Our last, best hope for a fantastic future. > .i ko na cpedu lo nu stidi vau loi jbopre .i danfu lu na go''i li''u .e > lu go''i li''u .i ji''a go''i lu na''e go''i li''u .e lu go''i na''i li''u .e > lu no''e go''i li''u .e lu to''e go''i li''u .e lu lo mamta be do cu sofybakni > li''u > > -- > You received this message because you are subscribed to the Google Groups > "Puppet Users" group. > To post to this group, send email to puppet-users@googlegroups.com. > To unsubscribe from this group, send email to > puppet-users+unsubscribe@googlegroups.com. > For more options, visit this group at > http://groups.google.com/group/puppet-users?hl=en. > >-- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Peter Berghold
2012-Jul-02 19:04 UTC
Re: [Puppet Users] puppetmasterd continuously consuming high CPU, with many interrupts
> > On Mon, Jul 2, 2012 at 3:01 PM, Ashley Penney <apenney@gmail.com> wrote: > It might be totally unrelated but check for ksoftirqd and see if it''s > running with high CPU. The leap second the other day caused all my > puppetmasters to spike up to 100% CPU and other people had similar problems. >Glad somebody besides me noticed that. My Nagios alerts went bezerk during leap second including CPU alterts for puppet master. Peter L. Berghold Salty.Cowdawg@gmail.com http://blog.berghold.net “"Know where to find the information and how to use it - That''s the secret of success" - Albert Einstein<http://www.quotesdaddy.com/quote/1184641/albert-einstein/know-where-to-find-the-information-and-how-to-use> ” Get this email app! <http://www.wisestamp.com/apps/quotes?utm_source=extension&utm_medium=email&utm_term=quotes&utm_campaign=apps> Want a signature like mine? <http://r1.wisestamp.com/r/landing?promo=16&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_16> Click here.<http://r1.wisestamp.com/r/landing?promo=16&dest=http%3A%2F%2Fwww.wisestamp.com%2Femail-install%3Futm_source%3Dextension%26utm_medium%3Demail%26utm_campaign%3Dpromo_16> -- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
llowder@oreillyauto.com
2012-Jul-02 19:06 UTC
[Puppet Users] Re: puppetmasterd continuously consuming high CPU, with many interrupts
On Monday, July 2, 2012 1:42:37 PM UTC-5, Robin Powell wrote:> > So, I have a server at home that has four VMs running inside it. > All are managed via puppet. The physical host runs puppetmasterd. > > I don''t recall noticing this before, but puppetmasterd has decided > to be kind of crazy. Here''s the physical host with no puppetmasterd > running: > >If this started this weekend, it may be related to the leapsecond that was applied on June 30 at midnight UTC. A restart should clear it up, or you can try the items listed here: http://serverfault.com/q/403732/121905 Lee -- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To view this discussion on the web visit https://groups.google.com/d/msg/puppet-users/-/g9Xdb5QlBTcJ. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Robin Lee Powell
2012-Jul-02 19:18 UTC
Re: [Puppet Users] Re: puppetmasterd continuously consuming high CPU, with many interrupts
On Mon, Jul 02, 2012 at 12:06:19PM -0700, llowder@oreillyauto.com wrote:> > > On Monday, July 2, 2012 1:42:37 PM UTC-5, Robin Powell wrote: > > > > So, I have a server at home that has four VMs running inside it. > > All are managed via puppet. The physical host runs > > puppetmasterd. > > > > I don''t recall noticing this before, but puppetmasterd has > > decided to be kind of crazy. Here''s the physical host with no > > puppetmasterd running: > > > > > If this started this weekend, it may be related to the leapsecond > that was applied on June 30 at midnight UTC. > > A restart should clear it up, or you can try the items listed > here: > > http://serverfault.com/q/403732/121905Turns out yes, it''s the leap second, but boy was the fix I found easier than that: http://artipc10.vub.ac.be/wordpress/2012/07/01/leap-second-causing-ksoftirqd-and-java-to-use-lots-of-cpu-time/ $ sudo date -s "`date`" Cleared it rigt up. -Robin -- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Ken Barber
2012-Jul-02 19:23 UTC
Re: [Puppet Users] Re: puppetmasterd continuously consuming high CPU, with many interrupts
> Turns out yes, it''s the leap second, but boy was the fix I found > easier than that: > > http://artipc10.vub.ac.be/wordpress/2012/07/01/leap-second-causing-ksoftirqd-and-java-to-use-lots-of-cpu-time/ > > $ sudo date -s "`date`" > > Cleared it rigt up.Huh. What a weird fix :-). ken. -- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Peter Brown
2012-Jul-03 00:19 UTC
Re: [Puppet Users] Re: puppetmasterd continuously consuming high CPU, with many interrupts
Thanks guys. I have been having a bizzare problem with java since then and this fixed it. On 3 July 2012 05:23, Ken Barber <ken@puppetlabs.com> wrote:>> Turns out yes, it''s the leap second, but boy was the fix I found >> easier than that: >> >> http://artipc10.vub.ac.be/wordpress/2012/07/01/leap-second-causing-ksoftirqd-and-java-to-use-lots-of-cpu-time/ >> >> $ sudo date -s "`date`" >> >> Cleared it rigt up. > > Huh. What a weird fix :-). > > ken. > > -- > You received this message because you are subscribed to the Google Groups "Puppet Users" group. > To post to this group, send email to puppet-users@googlegroups.com. > To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en. >-- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.
Walter Heck
2012-Jul-03 08:13 UTC
Re: [Puppet Users] Re: puppetmasterd continuously consuming high CPU, with many interrupts
The fix is actually not that weird, a call to clock_settime() function is enough (which setting the date on the command line does). More explained in the comments here: http://blog.mozilla.org/it/2012/06/30/mysql-and-the-leap-second-high-cpu-and-the-fix/ Walter On Tue, Jul 3, 2012 at 3:23 AM, Ken Barber <ken@puppetlabs.com> wrote:>> Turns out yes, it''s the leap second, but boy was the fix I found >> easier than that: >> >> http://artipc10.vub.ac.be/wordpress/2012/07/01/leap-second-causing-ksoftirqd-and-java-to-use-lots-of-cpu-time/ >> >> $ sudo date -s "`date`" >> >> Cleared it rigt up. > > Huh. What a weird fix :-). > > ken. > > -- > You received this message because you are subscribed to the Google Groups "Puppet Users" group. > To post to this group, send email to puppet-users@googlegroups.com. > To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en. >-- Walter Heck -- Check out my startup: Puppet training and consulting @ http://www.olindata.com Follow @olindata on Twitter and/or ''Like'' our Facebook page at http://www.facebook.com/olindata -- You received this message because you are subscribed to the Google Groups "Puppet Users" group. To post to this group, send email to puppet-users@googlegroups.com. To unsubscribe from this group, send email to puppet-users+unsubscribe@googlegroups.com. For more options, visit this group at http://groups.google.com/group/puppet-users?hl=en.