Recently I get a lot of dying apache childs, first only in dev mode,
now also in prod mode.
I'm using latest 1.2.8-DEV with propel on two debian machines.
The problem is the same on both.
On 23 Jun., 18:22, Bernhard Schussek <bschus...@gmail.com> wrote:
> Hi,
> Do you have XDebug installed? I found that XDebug sometimes gives
> useful error messages when plain PHP segfaults.
Just installed it - but it shows nothing, just 3 segfaults per
request...
[Wed Jun 24 09:32:36 2009] [notice] child pid 9081 exit signal
Segmentation fault (11)
[Wed Jun 24 09:32:37 2009] [notice] child pid 9082 exit signal
Segmentation fault (11)
[Wed Jun 24 09:32:37 2009] [notice] child pid 9083 exit signal
Segmentation fault (11)
So I used the old way of inserting a lot of echo's and found out, that
inserting a ->__toString() cured this segfault.
(I gave an object to a function wanting a string)
- but we were unable to create a short example script - it seems
whenever we removed something from our larger symfony project, to
isolate the segfault, the segfault would disappear
Casting the object to a string in a parent function does not produce a
segfault, while calling a function and casting in the called function
does. Even passing the object as the first as opposed to fourth
argument of the function causes the segfault to disappear.
We're not using XCache or Memcache
On Jun 24, 3:37 am, skr <konrad.rie...@googlemail.com> wrote:
> On 23 Jun., 18:22, Bernhard Schussek <bschus...@gmail.com> wrote:> Hi,
> > Do you have XDebug installed? I found that XDebug sometimes gives
> > useful error messages when plain PHP segfaults.
> Just installed it - but it shows nothing, just 3 segfaults per
> request...
> [Wed Jun 24 09:32:36 2009] [notice] child pid 9081 exit signal
> Segmentation fault (11)
> [Wed Jun 24 09:32:37 2009] [notice] child pid 9082 exit signal
> Segmentation fault (11)
> [Wed Jun 24 09:32:37 2009] [notice] child pid 9083 exit signal
> Segmentation fault (11)
> So I used the old way of inserting a lot of echo's and found out, that
> inserting a ->__toString() cured this segfault.
> (I gave an object to a function wanting a string)
What DBMS are you using? some versions of php's postgresql extensions
are known to segfault systematically, at a very late stage of php
process lifetime (ie after everything is done).
One simple symptom for this, if you use the same php and php
extensions from cli, is to look for a segfault at the end of every
database cli task.
Anyway, as some pages are ok, it's maybe not this.
Interesting topic indeed, it's pure experience sharing.
> Recently I get a lot of dying apache childs, first only in dev mode,
> now also in prod mode.
> I'm using latest 1.2.8-DEV with propel on two debian machines.
> The problem is the same on both.
> What DBMS are you using? some versions of php's postgresql extensions
> are known to segfault systematically, at a very late stage of php
> process lifetime (ie after everything is done).
Ouch, happy to hear that, I didn't find much information about this
problem that I have very often.
Are there workarounds for this ? Any links where I can find
informations about ?
I have the same problem than Romain. I'm using the last symfony
version with postgres 8.3 and doctrine. Think i read the problem could
be the encoding or sh related to that, can it be possible??.
On 26 jun, 14:03, Tonio <brankg...@gmail.com> wrote:
> > What DBMS are you using? some versions of php's postgresql extensions
> > are known to segfault systematically, at a very late stage of php
> > process lifetime (ie after everything is done).
> Ouch, happy to hear that, I didn't find much information about this
> problem that I have very often.
> Are there workarounds for this ? Any links where I can find
> informations about ?
|
