[rjb-users] Daemonized mongrel causes Segmentation faults

Andrès Koetsier koetsier at oblivion.nl
Tue Nov 20 04:41:57 EST 2007


Some information about the system:

ruby 1.8.5 (2006-08-25) [x86_64-linux]
Java(TM) SE Runtime Environment (build 1.6.0_03-b05)
mongrel_cluster 1.0.5
mongrel 1.1.1
daemons 1.0.7
rjb 1.0.11
rails 1.2.3
Linux 2.6.9 (Redhat)

Regards

rjb-users at rubyforge.org wrote:
> Hi,
>
> Now I'm trying to reproduce the problem, but in my environment,
> daemon mongrel runs fine with rjb, so please inform me more details.
> With my test environment is
> ruby 1.8.6-pl110 x86 64-linux
> Java build 1.6.0-b105
>  with poi-2.5.1.jar (my rjb testing app is to produce excel files with
> Jakarta -POI, but I suppose that the Java libraries are not concerned)
> mongrel 1.1.1
> daemons 1.0.9
> rjb 1.0.11
> rails 1.2.3
> Linux 2.6.20 (Ubuntu 4.1.2)
>
> Regards
>
> On Thu, 15 Nov 2007 20:36:20 +0100
> rjb-users at rubyforge.orgさん wrote:
>
>   
>> Hello,
>>
>> I already read that blog and used the memory configuration flags in the
>> Rjb::load command, so the problem lies somewhere else.... Still, it is
>> quite strange mongrel only crashes when being run as a daemon. It cannot
>> be a user/right problem since I tried starting mongrel as the mongrel
>> user, which works nicely without the -d.
>>
>> Regards
>>
>> rjb-users at rubyforge.org wrote:
>>     
>>> Hi,
>>>
>>> Then, have you already read http://jaywhy.wordpress.com/2007/03/05/pdf-templates-via-rails/ ?
>>> If you've not, then please read it.
>>> I suppose that you can either set the options -Xms... in the code as
>>> Rjb::load('', ['-Xms16m', '-Xmx32m'])
>>>
>>> Regards
>>>
>>> On Thu, 15 Nov 2007 16:46:15 +0100
>>> rjb-users at rubyforge.orgさん wrote:
>>>
>>>
>>>       
>>>> I already tried that solution. Procfs is mounted correctly and it is
>>>> only a problem when running mongrel as a daemon. I recon if procfs was
>>>> not mounted (or incorrectly) it should also have segfaulted in 'normal'
>>>> mode.
>>>>
>>>> Regards
>>>>
>>>> rjb-users at rubyforge.org wrote:
>>>>
>>>>         
>>>>> Hmmmm, interesting.
>>>>>
>>>>> http://forum.java.sun.com/thread.jspa?threadID=542672&messageID=2634141
>>>>> I'm not sure the forum thread's result - just mounting linprocfs for you,
>>>>> but please try it first.
>>>>> And please let me know the result.
>>>>>
>>>>> Regards
>>>>>
>>>>> On Thu, 15 Nov 2007 12:24:43 +0100
>>>>> rjb-users at rubyforge.orgさん wrote:
>>>>>
>>>>>
>>>>>
>>>>>           
>>>>>> Hello,
>>>>>>
>>>>>> I am currently trying to generate PDF's from ruby using iText. On my
>>>>>> development machine everything worked fine. However when I moved to the
>>>>>> production server (which uses mongrel_cluster), every instance of
>>>>>> mongrel crashes when initializing the Java VM.
>>>>>>
>>>>>> When I started a server manually with 'sudo -u mongrel ruby
>>>>>> script/server ......', generating PDF's worked again. However when I
>>>>>> start the server as a daemon 'sudo -u mongrel ruby script/server -d
>>>>>> ......' the instance crashes again with a segmentation fault. The line
>>>>>> before the segmentation fault reads 'Java HotSpot(TM) 64-Bit Server VM
>>>>>> warning: Can't detect initial thread stack location - find_vma failed'.
>>>>>> The machine is a Intel Xeon system.
>>>>>>
>>>>>> Is this a known bug, and if so is there already a solution?
>>>>>>
>>>>>> Thanks in advance,
>>>>>> Regards Andres
>>>>>> _______________________________________________
>>>>>> rjb-users mailing list
>>>>>> rjb-users at rubyforge.org
>>>>>> http://rubyforge.org/mailman/listinfo/rjb-users
>>>>>>
>>>>>>
>>>>>>             
>>>>> --
>>>>> arton
>>>>>
>>>>>
>>>>>
>>>>>           
>>> --
>>> arton
>>>
>>>
>>>       
>
> --
> arton
>
>   



More information about the rjb-users mailing list