Re: [squid-users] running out of filedescriptors

James Camaron wrote:
>
> Hello Marc,
>
> My cache.log shows as follows
>
 
 Yes , but I forgot that you were doing tests with
 polygraph.

 Does this test represent your intended load , because
 polygraph may need extreme setting which you may not
 need during normal usage.

 Normally SQUID will be happy with the intended number
 of file desc's.
  
 Unless you can 'justify' that polygraph conditions,
 will also hit your SQUID in 'real life'

 M.

> 2002/09/17 17:55:21| Starting Squid Cache version
> 2.5.STABLE1 for hppa2.0w-hp-hp
> ux11.11...
> 2002/09/17 17:55:21| Process ID 1964
> 2002/09/17 17:55:21| With 16384 file descriptors
> available
> 2002/09/17 17:55:21| Performing DNS Tests...
> 2002/09/17 17:55:21| Successful DNS name lookup
> tests...
> 2002/09/17 17:55:21| DNS Socket created at 0.0.0.0,
> port 49236, FD 5
> 2002/09/17 17:55:21| Adding nameserver 172.16.1.200
> from /etc/resolv.conf
> 2002/09/17 17:55:21| Unlinkd pipe opened on FD 10
> 2002/09/17 17:55:21| Swap maxSize 0 KB, estimated 0
> objects
> 2002/09/17 17:55:21| Target number of buckets: 0
> 2002/09/17 17:55:21| Using 8192 Store buckets
> 2002/09/17 17:55:21| Max Mem size: 512000 KB
> 2002/09/17 17:55:21| Max Swap size: 0 KB
> 2002/09/17 17:55:21| Using Least Load store dir
> selection
> 2002/09/17 17:55:21| Set Current Directory to
> /usr/local/squid/var/cache
> 2002/09/17 17:55:21| Loaded Icons.
> 2002/09/17 17:55:21| Accepting HTTP connections at
> 0.0.0.0, port 3128, FD 9.
> 2002/09/17 17:55:21| Accepting ICP messages at
> 0.0.0.0, port 3130, FD 11.
> 2002/09/17 17:55:21| WCCP Disabled.
> 2002/09/17 17:55:21| Ready to serve requests.
> 2002/09/17 17:55:21| Finished rebuilding storage from
> disk.
> 2002/09/17 17:55:21| 0 Entries scanned
> 2002/09/17 17:55:21| 0 Invalid entries.
> 2002/09/17 17:55:21| 0 With invalid flags.
> 2002/09/17 17:55:21| 0 Objects loaded.
> 2002/09/17 17:55:21| 0 Objects expired.
> 2002/09/17 17:55:21| 0 Objects cancelled.
> 2002/09/17 17:55:21| 0 Duplicate URLs purged.
> 2002/09/17 17:55:21| 0 Swapfile clashes
> avoided.
> 2002/09/17 17:55:21| Took 0.0 seconds ( 0.0
> objects/sec).
> 2002/09/17 17:55:21| Beginning Validation Procedure
> 2002/09/17 17:55:22| Completed Validation Procedure
> 2002/09/17 17:55:22| Validated 0 Entries
> 2002/09/17 17:55:22| store_swap_size = 0k
> 2002/09/17 17:55:22| storeLateRelease: released 0
> objects
> 2002/09/17 17:56:58| httpAccept: FD 9: accept failure:
> (23) File table overflow
> 2002/09/17 17:56:58| comm_open: socket failure: (23)
> File table overflow
> 2002/09/17 17:56:58| Reserved FD adjusted from 100 to
> 15982 due to failures
> 2002/09/17 17:56:58| WARNING! Your cache is running
> out of filedescriptors
> 2002/09/17 17:57:14| WARNING! Your cache is running
> out of filedescriptors
> 2002/09/17 17:57:30| WARNING! Your cache is running
> out of filedescriptors
>
> Regards,
> James
>
> --- Marc Elsen <marc.elsen@imec.be> wrote:
> >
> >
> > James Camaron wrote:
> > >
> > > Hello All,
> > >
> > > I have installed squid in HP-UX. Then i start
> > pumping
> > > request to squid using polygraph.
> > >
> > > I gave 300 req/sec. In the cache.log i get the
> > > following warning continuously
> > >
> > > WARNING! Your cache is running out of
> > filedescriptors
> > >
> > > I already increased the maxfiles to 16384.
> > >
> > > Where i'm going wrong.
> > >
> > Nowhere, the question is, does SQUID now use the
> > intended
> > number of file desc's.
> > (What does SQUID say on startup in cache.log ?).
> >
> > M.
> >
> > > Regards,
> > > James
> > >
> > > __________________________________________________
> > > Do you Yahoo!?
> > > Faith Hill - Exclusive Performances, Videos & More
> > > http://faith.yahoo.com
> >
> > --
> >
> > 'Time is a consequence of Matter thus
> > General Relativity is a direct consequence of QM
> > (M.E. Mar 2002)
>
> __________________________________________________
> Do you Yahoo!?
> Faith Hill - Exclusive Performances, Videos & More
> http://faith.yahoo.com

-- 
 'Time is a consequence of Matter thus
 General Relativity is a direct consequence of QM
 (M.E. Mar 2002)