Re: [squid-users] Re: 1.2 problems on Solaris v2.5 (ss20)

On Thu, Sep 17, 1998 at 01:15:53PM +0400, Alexander Serkin wrote:

> Alex Rousskov wrote:
> >
> > What patches did you apply to b24 if any?
>
> none, unfortunately. Do you mean first patch set ?

Ack Ack Ack. Apply the first patch set to 1.2b24 under Solaris + async-io,
under _all_ circumstances!

If you don't use any special malloc handling (ie, c-library routines), then
the process will periodically catch BUS errors and abort. It won't dump core
though, it'll stall whilst trying to write the swap log file. The only way
to kill it is with a "-9" signal (although I didn't try using the squid -k
option).

If you use dlmalloc, then it'll catch a segmentation violation and die with
a core dump immediately upon starting up.

Both these problems go away with the first patch set applied, or if you
don't enable asynchronous I/O. I have 1.2b24+first patch set + gnumalloc on
Solaris 2.6 running very nicely indeed. I've hit other bugs with it, but I
can live with them.

I especially like the _much_ reduced memory footprint + the vastly increased
performance :-)

Cheers,

Chris

-- 
Chris Tilbury, UNIX Systems Administrator, IT Services, University of Warwick
EMAIL: cudch+s@csv.warwick.ac.uk PHONE: +44 1203 523365(V)/+44 1203 523267(F)
                            URL: http://www.warwick.ac.uk/staff/Chris.Tilbury