Re: [squid-users] squid run with 100% usage

What "unixtime bug"?

--
Henrik
Dennis Berger wrote:
> 
> I recently answered myself ...
> I build a new squid the old one had the unixtime bug.
> 
> Henrik Nordstrom schrieb:
> 
> > Hard to tell only from this.
> >
> > What you need to check is what fd 10 (or whatever FD is repeatedly
> > reported as ready by poll) being used for when Squid is running at 100%
> > CPU?
> >
> > Now some configuration questions:
> >
> > a) Are you using --enable-time-hack?
> >
> > b) Are you using delay pools?
> >
> > c) Is there any suspecible requests in access.log at the time (or
> > actually, at the time when things settles down to normal again...)
> >
> > d) What type of cache directory are you using?
> >
> > And please try upgrading to a current Squid release..
> >
> > Regards
> > Henrik Nordström
> > Squid Hacker
> >
> > Dennis Berger wrote:
> > >
> > > Hi my squid shows 100% usage in top an strace shows me this messages.
> > > Btw the squid.dump raises to 5 mb in 2 seconds, with millions of the
> > > same message...
> > > -------------------strace -o /tmp/squid.dump -p 345-----------------
> > > gettimeofday({1002901170, 457031}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 457186}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 457341}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 457496}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 457667}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 457823}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 457977}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 458132}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 458287}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 458442}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 458596}, NULL) = 0
> > > poll([{fd=8, events=POLLIN}, {fd=9, events=POLLIN}, {fd=10,
> > > events=POLLIN}], 3, 0) = 0
> > > gettimeofday({1002901170, 458750}, NULL) = 0
> > > gettimeofday({1002901170, 458810}, NULL) = 0
> > > -----------------------------------------------------------------
> > >
> > > Its version squid-2.3.STABLE4-10.6.2.i386.rpm
> > > Can somebody explain what's going on here ?
> > > -----------------------------TOP--------------------
> > >   5:42pm  up  1:05,  1 user,  load average: 0.56, 0.50, 0.50
> > > 41 processes: 38 sleeping, 3 running, 0 zombie, 0 stopped
> > > CPU states: 40.0% user, 35.2% system,  0.0% nice, 24.6% idle
> > > Mem:   127764K av,  120708K used,    7056K free,    7792K shrd,   48592K
> > >
> > > buff
> > > Swap:  208056K av,    5844K used,  202212K free                   49592K
> > >
> > > cached
> > >
> > >   PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
> > >  1420 squid     19   0  9624 9624  1144 R    99.0  7.5   3:05 squid
> > > -------------------------------------------------