[phpa] Re: Reliability and uptime
- From: "Geoff Caplan" <geoff@xxxxxxxxxxxx>
- To: <phpa@xxxxxxxxxxxxx>
- Date: Fri, 28 Dec 2001 23:48:02 -0000
Hi folks
Thanks for the detailed feedback - it gives me the confidence to give phpa a
try. I think I'll wait for the final release of 1.2 and then get down to
some testing.
Thanks again
Geoff Caplan
Advantae Ltd.
----- Original Message -----
From: "Nick Lindridge" <nick@xxxxxxxxxxxxxxxxx>
To: <phpa@xxxxxxxxxxxxx>
Sent: Friday, December 28, 2001 5:42 PM
Subject: [phpa] Re: Reliability and uptime
>
> Hi Geoff,
>
> On Fri, Dec 28, 2001 at 12:41:56PM -0000, Geoff Caplan wrote:
> >
> > Nick
> >
> > I have an application that involves some heavy lifting ( with some
requests
> > running over 4000 lines of code) so finding the right cache solution is
a
> > priority.
> >
> > But my Linux admin skills are limited so I really need a "set and
forget"
> > solution. How confident are you that 1.2 will be relatively problem-free
and
> > will allow long periods of uptime on RedHat 7.1?
> >
> > Lurking on their lists, it seems that some of the other non-commercial
> > solutions seem to be struggling to achieve this kind of reliabliity. Are
you
> > pretty confident that you have cracked it? Is it safe and sensible for a
> > Linux newbie to deploy phpa on a production server?
>
> Thanks for your interest in PHPA. Regarding stability, it's likely that
you
> should be fine, but I would highly recommend a period (days rather than
hours)
> of testing for your own particular setup. Any problems of memory
corruption
> within PHP itself may be more likely to show up with PHPA installed
because
> memory is reused from one request to the next rather than effectively
being
> reinitialised for each request, and it's possible that there are obscure
> bugs that haven't yet been exposed within PHPA itself.
>
> I am aware of one site that has occasional problems, but the cause isn't
> identified, and from what I can tell, most sites that try it work
successfully.
>
> I would also suggest that you use at least 4.0.7RC3 as this fixes
> some memory corrupting bugs in 4.0.6
>
> Version 1.2 is far less susceptible to locking up apache if apache
> crashes because phpa now doesn't lock while executing scripts, only
> while locating them in the shm cache or writing to the cache. It also
> attempts to detect processes that have crashed and releases any locks
> that were held at the point of crashing. However this mechanism isn't
> foolproof.
>
> Try it and monitor it for a bit and see how you get on, and turn on
> the shm logging and make sure you check the apache errors log for
> signs of problems.
>
> Hope this helps,
>
> Nick
>
> ------------------------------------------------------------------------
> www.php-accelerator.co.uk Home of the free PHP Accelerator
>
> To post, send email to phpa@xxxxxxxxxxxxx
> To unsubscribe, email phpa-request@xxxxxxxxxxxxx with subject unsubscribe
>
>
------------------------------------------------------------------------
www.php-accelerator.co.uk Home of the free PHP Accelerator
To post, send email to phpa@xxxxxxxxxxxxx
To unsubscribe, email phpa-request@xxxxxxxxxxxxx with subject unsubscribe
- Follow-Ups:
- [phpa] Re: Reliability and uptime
- From: Nick Lindridge
- References:
- [phpa] Re: Accelerator or Caching?
- From: Anthony Walker
- [phpa] Solaris 2.8 version
- From: Manuel Lemos
- [phpa] Re: Solaris 2.8 version
- From: Nick Lindridge
- [phpa] Reliability and uptime
- From: Geoff Caplan
- [phpa] Re: Reliability and uptime
- From: Nick Lindridge
Other related posts:
- » [phpa] Reliability and uptime
- » [phpa] Re: Reliability and uptime
- » [phpa] Re: Reliability and uptime
- » [phpa] Re: Reliability and uptime
- » [phpa] Re: Reliability and uptime
- » [phpa] Re: Reliability and uptime
- [phpa] Re: Reliability and uptime
- From: Nick Lindridge
- [phpa] Re: Accelerator or Caching?
- From: Anthony Walker
- [phpa] Solaris 2.8 version
- From: Manuel Lemos
- [phpa] Re: Solaris 2.8 version
- From: Nick Lindridge
- [phpa] Reliability and uptime
- From: Geoff Caplan
- [phpa] Re: Reliability and uptime
- From: Nick Lindridge