500

Sébastien Aperghis-Tramoni sebastien at aperghis.net
Wed Jul 31 11:56:23 CEST 2013


Andrew Shitov a écrit :

> Hi guys,

Hello Andrew,

> Yesterday there was a number of 500 internal server error on the
> ye2013 Act instance. I just got one again today. Is it a problem on
> the server or increased load or such? Could someone please take
> a look at this?

There doesn't seem to be an increased load. The only difference is 
that since saturday, the server is running using the code from the
github repository, branch master. It seem that some of the included
changes have the effect to make Apache segfaults pretty hard:

[Wed Jul 31 11:09:27 2013] [notice] child pid 20883 exit signal Bus error (10)
[Wed Jul 31 11:09:38 2013] [notice] child pid 20845 exit signal Bus error (10)
[Wed Jul 31 11:09:49 2013] [notice] child pid 20840 exit signal Segmentation fault (11)
[Wed Jul 31 11:09:54 2013] [notice] child pid 20849 exit signal Bus error (10)
Bad free() ignored (PERL_CORE) at /usr/local/lib/perl5/site_perl/5.8.9/Wiki/Toolkit/Store/Database.pm line 118.
[Wed Jul 31 11:09:55 2013] [error] Bizarre copy of UNKNOWN in helem at /usr/local/lib/perl5/site_perl/5.8.9/Wiki/Toolkit.pm line 141.\n

then it regularly generates these errors:

[Wed Jul 31 11:10:55 2013] [error] Bizarre copy of UNKNOWN in sassign at /usr/local/lib/perl5/site_perl/5.8.9/mach/Apache/Request.pm line 36.\n

Like yesterday, I restarted Apache, but it doesn't look very promising.
I guess we'll have to somehow patch the wiki part so it doesn't make
Apache crash again.

-- 
Sébastien Aperghis-Tramoni

Close the world, txEn eht nepO.


More information about the Act mailing list