View previous topic :: View next topic |
Author |
Message |
MiXi-IL n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
Joined: 21 Aug 2005 Posts: 36 Location: Technion
|
Posted: Tue Mar 21, 2006 10:26 pm Post subject: dhcpcd takes 1%-3% constant cpu resources out of my system |
|
|
Hi,
I used top today and found out that dhcpcd is using my cpu constantly, between 1%-3% utilization.
Is it normal for such a process to consume such amounts of cpu time (constantly...)?
What can I do to return to normal state where I get to see 100% idle once in a while (or at least 99... )
Thanks. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
UberLord Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/16007251014200867ea775c.gif)
Joined: 18 Sep 2003 Posts: 6835 Location: Blighty
|
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
mliesenf n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
Joined: 22 Jul 2003 Posts: 18 Location: University of Florida
|
Posted: Wed Apr 05, 2006 3:56 am Post subject: |
|
|
I'm noticing something very similar. /sbin/dhcpcd seems to use more CPU time when I'm actively scp'ing a large set of files. I can't explain why, but that's just something I'm noticing.
dhcpcd version: net-misc/dhcpcd-2.0.3 |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
MiXi-IL n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
Joined: 21 Aug 2005 Posts: 36 Location: Technion
|
Posted: Wed Apr 05, 2006 5:28 pm Post subject: |
|
|
It seems that I missed something when I diagnosed the situation. Now when this thread has woke up again, I should correct it.
When I executed "top", I didn't notice that on another tty I had ethereal running. After I terminated ethereal (a few days later) dhcpcd has returned to normal cpu usage. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
gregp01 Tux's lil' helper
![Tux's lil' helper Tux's lil' helper](/images/ranks/rank_rect_1.gif)
Joined: 23 Feb 2004 Posts: 103
|
Posted: Sun Apr 30, 2006 9:19 pm Post subject: |
|
|
mliesenf / anyone: Have you had any luck figuring out what the problem was?
I'm experiencing the same issue with dhcpcd using a constant 1% - 3% CPU. Mine seems to be related to Samba (smbd) activity rather than scp, though. Regardless, it seems to me that this should *never* happen - dhcpcd should take much less than a second of CPU time every few hours, when it renews the DHCP lease. On my server with an uptime of several days, top reports over 4 hours of CPU time! I too am using dhcpcd-2.0.3. It seems like this could be a bug, as I never saw this behavior before. I'm going to try reverting to the previous version and see what happens... |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
gregp01 Tux's lil' helper
![Tux's lil' helper Tux's lil' helper](/images/ranks/rank_rect_1.gif)
Joined: 23 Feb 2004 Posts: 103
|
Posted: Sat May 13, 2006 10:16 pm Post subject: |
|
|
Well, I've now tried versions 2.0.4, 2.0.5, and reverting to 2.0.0. The issue is still present in 2.0.4, and to a lesser degree in 2.0.5 (occasional jumps to a few fractions of a percent of CPU), and is completely absent from 2.0.0. I'm going to stick with 2.0.0 for now, and I've submitted a bug report. Hopefully someone will be able to track this down. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
UberLord Retired Dev
![Retired Dev Retired Dev](/images/ranks/rank-retired.gif)
![](images/avatars/16007251014200867ea775c.gif)
Joined: 18 Sep 2003 Posts: 6835 Location: Blighty
|
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
no_hope Guru
![Guru Guru](/images/ranks/rank_rect_3.gif)
![](images/avatars/17548115346b0bdac11ddf.jpg)
Joined: 23 Jun 2003 Posts: 482
|
Posted: Wed May 17, 2006 2:23 pm Post subject: |
|
|
UberLord wrote: | Care to profile each version so we can find out where it starts to go wrong? |
how would I go about doing that?
Below is a taste of what strace -f dhcpcd eth0 gives me. The select on the first line is where the child sits and waits after it a lease has been acquired. The rest of the snippet is what happens continuously as I run my scripts on a Samba mount. During that time dhcpcd eats ~5% of CPU time on my Samba server and 10-20% of CPU time on the client. I am using dhcpcd 2.0.3 on both ends, strace was run on the client.
Code: |
select(5, [3 4], NULL, NULL, NULL) = 1 (in [3])
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0\263H\6@\0@\6\334"..., 512, 0, NULL, NULL) = 193
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\0\331\5\374@\0@\6\36"..., 512, 0, NULL, NULL) = 231
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0004H\7@\0@\6\334\260"..., 512, 0, NULL, NULL) = 66
recvfrom(3, 0xbfb2614c, 512, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
select(5, [3 4], NULL, NULL, NULL) = 1 (in [3])
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0\221H\10@\0@\6\334"..., 512, 0, NULL, NULL) = 159
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\0i\5\376@\0@\6\36\205"..., 512, 0, NULL, NULL) = 119
recvfrom(3, 0xbfb2614c, 512, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
select(5, [3 4], NULL, NULL, NULL) = 1 (in [3])
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0aH\t@\0@\6\334\201"..., 512, 0, NULL, NULL) = 111
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\0[\6\0@\0@\6\36\221"..., 512, 0, NULL, NULL) = 105
recvfrom(3, 0xbfb2614c, 512, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
select(5, [3 4], NULL, NULL, NULL) = 1 (in [3])
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0\243H\n@\0@\6\334"..., 512, 0, NULL, NULL) = 177
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\0\331\6\2@\0@\6\36"..., 512, 0, NULL, NULL) = 231
recvfrom(3, 0xbfb2614c, 512, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
select(5, [3 4], NULL, NULL, NULL) = 1 (in [3])
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0\221H\v@\0@\6\334"..., 512, 0, NULL, NULL) = 159
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\0i\6\4@\0@\6\36\177"..., 512, 0, NULL, NULL) = 119
recvfrom(3, 0xbfb2614c, 512, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
select(5, [3 4], NULL, NULL, NULL) = 1 (in [3])
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0\247H\f@\0@\6\334"..., 512, 0, NULL, NULL) = 181
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\0\331\6\6@\0@\6\36"..., 512, 0, NULL, NULL) = 231
recvfrom(3, 0xbfb2614c, 512, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
select(5, [3 4], NULL, NULL, NULL) = 1 (in [3])
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0\257H\r@\0@\6\334"..., 512, 0, NULL, NULL) = 189
recvfrom(3, 0xbfb2614c, 512, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
select(5, [3 4], NULL, NULL, NULL) = 1 (in [3])
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\334\6\10@\0@\6\31"..., 512, 0, NULL, NULL) = 512
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\334\6\n@\0@\6\31"..., 512, 0, NULL, NULL) = 512
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0004H\16@\0@\6\334"..., 512, 0, NULL, NULL) = 66
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\334\6\f@\0@\6\31"..., 512, 0, NULL, NULL) = 512
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\334\6\16@\0@\6\31"..., 512, 0, NULL, NULL) = 512
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0004H\17@\0@\6\334"..., 512, 0, NULL, NULL) = 66
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\334\6\20@\0@\6\31"..., 512, 0, NULL, NULL) = 512
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\334\6\22@\0@\6\30"..., 512, 0, NULL, NULL) = 512
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0004H\20@\0@\6\334"..., 512, 0, NULL, NULL) = 66
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\334\6\24@\0@\6\30"..., 512, 0, NULL, NULL) = 512
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\334\6\26@\0@\6\30"..., 512, 0, NULL, NULL) = 512
recvfrom(3, "\0\260\320\222Un\0\f)\2y\27\10\0E\0\0004H\21@\0@\6\334"..., 512, 0, NULL, NULL) = 66
recvfrom(3, "\0\f)\2y\27\0\260\320\222Un\10\0E\0\5\304\6\30@\0@\6\31"..., 512, 0, NULL, NULL) = 512
recvfrom(3, 0xbfb2614c, 512, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable)
|
|
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
bobwarlock n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
Joined: 17 Sep 2003 Posts: 22 Location: Lincoln, NE
|
Posted: Fri Jul 21, 2006 10:33 pm Post subject: Fixed in dhcpcd-2.0.6 |
|
|
Just in case people miss/ignore the bug report mentioned above, this has been resolved in version 2.0.6 and up.
I was getting ~15-20% cpu usage while doing a large scp on my system, and upgrading to 2.0.8 solved it. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
turtles Veteran
![Veteran Veteran](/images/ranks/rank_rect_5_vet.gif)
![](images/avatars/8003843766096d82d4fbf9.png)
Joined: 31 Dec 2004 Posts: 1698
|
Posted: Sat Jul 22, 2006 4:32 am Post subject: |
|
|
side note: scp starts a process for each file, might try using tar and or sync for lots of files. _________________ Donate to Gentoo |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
gregp01 Tux's lil' helper
![Tux's lil' helper Tux's lil' helper](/images/ranks/rank_rect_1.gif)
Joined: 23 Feb 2004 Posts: 103
|
Posted: Wed Aug 16, 2006 12:50 pm Post subject: |
|
|
It took me forever to get another chance to look into this, but it definitely looks to have been fixed. Thanks a lot, everyone! |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
|
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|