[libvirt] the POLLHUP envet for a pipe descriptor is the cause that the libvirt_lxc costs 100% cpu usage.
8:04 p.m.
I found the poll in virEventPollRunOnce (vireventpoll.c), a POLLHUP event
occurred for a pipe descriptor, thus the poll always returned because of
it.( the pipe broken). this may be the best I could track temporarily，I
don't know what to do netx , to mask the hangup event or to do something to
handle the hangup event , what to do ? Is it a bug or something ?
any help would be appreciate. :)
4:56 a.m.
On Fri, Nov 01, 2013 at 09:04:09AM +0800, jj zhong wrote:
I found the poll in virEventPollRunOnce (vireventpoll.c), a POLLHUP
event
occurred for a pipe descriptor, thus the poll always returned because of
it.( the pipe broken). this may be the best I could track temporarily，I
don't know what to do netx , to mask the hangup event or to do something to
handle the hangup event , what to do ? Is it a bug or something ?
Obviously there is a bug, since we should never go 100% CPU burn. We are
already handling POLLHUP though - it gets mapped to VIR_EVENT_HANDLE_HANGUP
Not sure why we continue to burn cycles though.
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
6:37 a.m.
New subject: [libvirt] the POLLHUP envet for a pipe descriptor is the causethat the libvirt_lxc costs 100% cpu usage.
I put some debug info here https://bugzilla.redhat.com/show_bug.cgi?id=1005570
It seems something happened in virLXCControllerConsoleUpdateWatch function,
a pipe , the write side broken, so the read side hangup which causes 100% cpu usage.
On my compute , libvirt_lxc started, 3 hours laster, this would occur .
In early time , under normal situation , the debug would print informations on
newHostEvents or newContEvents ,
but then It's going to be crazy ... no debug information on newHostEvents or
newContEvents anymore ,just infinate loop.
I even could not understand what's the meaning of the code ~~
------------------ Original ------------------
From: "Daniel P. Berrange";<berrange(a)redhat.com>;
Date: Fri, Nov 1, 2013 05:56 PM
To: "jj zhong"<menghuizhanguo(a)gmail.com>;
Cc: "libvirt-list"<libvirt-list(a)redhat.com>;
Subject: Re: [libvirt] the POLLHUP envet for a pipe descriptor is the causethat the
libvirt_lxc costs 100% cpu usage.
On Fri, Nov 01, 2013 at 09:04:09AM +0800, jj zhong wrote:
I found the poll in virEventPollRunOnce (vireventpoll.c), a POLLHUP
event
occurred for a pipe descriptor, thus the poll always returned because of
it.( the pipe broken). this may be the best I could track temporarily，I
don't know what to do netx , to mask the hangup event or to do something to
handle the hangup event , what to do ? Is it a bug or something ?
Obviously there is a bug, since we should never go 100% CPU burn. We are
already handling POLLHUP though - it gets mapped to VIR_EVENT_HANDLE_HANGUP
Not sure why we continue to burn cycles though.
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
4073
days inactive
4073
days old
2 comments
3 participants
participants (3)
-
Daniel P. Berrange -
jj -
jj zhong