[erlang-questions] Investigate an infinite loop on production servers

Morgan Segalis msegalis@REDACTED
Fri May 24 10:13:46 CEST 2013


The problem is that the VM freezes completely, it does not generate a crash dump

Is there a way to limit the memory that a VM can allocate, so the server is not overwhelmed in order to create a crash dump ?

Le 24 mai 2013 à 02:00, Vance Shipley <vances@REDACTED> a écrit :

> Have you analyzed the crash dump file with the crash dump viewer?
> On May 24, 2013 3:00 AM, "Morgan Segalis" <msegalis@REDACTED> wrote:
> Yeah you got that right ! leaking at a huge rate at some point !
> 
> - The number of Fd - I don't get close to the max
> # cat /proc/sys/fs/file-nr
> 3264    0       6455368
> 
> - On the production server there is only the erlang node, no other service…
> The beam.smp was through the roof at 300% CPU and 97% RAM
> The weird thing is that it got there in a second, I was looking at it when it happens.
> 
> - It has happened with 2000 connections, 4000 connections, and 10000 connections… 5 min after start, 5hours after start.
> 
> I really can't find a pattern here…and I'm becoming a little desperate.
> 
> Thank you for your help again.
> 
> Morgan.
> 
> Le 23 mai 2013 à 23:20, Dmitry Kolesnikov <dmkolesnikov@REDACTED> a écrit :
> 
>> You system definitely leaking some resources :-/
>>  - Check number of used FD(s) may be you exceeded limit there 
>>  - What was overall system memory / cpu utilisation before crash?
>>  - Check how many connections you got before crash, may be you can reproduce it at dev
>> 
>> - Dmitry
>> 
>> On May 24, 2013, at 12:13 AM, Morgan Segalis <msegalis@REDACTED> wrote:
>> 
>>> Ok, it finally got into the infinite loop…
>>> 
>>> And of course, the node on which I was running etop could not give me anymore since it got disconnected from the production node.
>>> 
>>> So back to square one… no way to investigate correctly so far :-/
>>> 
>>> Morgan.
>>> 
>>> Le 23 mai 2013 à 16:34, Morgan Segalis <msegalis@REDACTED> a écrit :
>>> 
>>>> Yeah that what I'm doing right now, but of course, when I'm monitoring it, it won't crash, only when I sleep !!
>>>> 
>>>> I get back to the Erlang list as soon as I have more informations about this.
>>>> 
>>>> Thank you all !
>>>> 
>>>> Morgan.
>>>> 
>>>> Le 23 mai 2013 à 16:30, Vance Shipley <vances@REDACTED> a écrit :
>>>> 
>>>>> Keep etop running and capture the output to a file (e.g. etop ... | tee stop.log). After it gets into trouble look back and see what was happening beforehand.
>>>>> On May 23, 2013 6:16 PM, "Morgan Segalis" <msegalis@REDACTED> wrote:
>>>>> So I should go back to R15B ?
>>>>> 
>>>>> erlang:memory() gives me 
>>>>> 
>>>>> [{total,1525779584},
>>>>>  {processes,1272881427},
>>>>>  {processes_used,1272789743},
>>>>>  {system,252898157},
>>>>>  {atom,372217},
>>>>>  {atom_used,346096},
>>>>>  {binary,148093608},
>>>>>  {code,8274446},
>>>>>  {ets,1546832}]
>>>>> 
>>>>> 
>>>>> But keep in mind that right now, there is no infinite loop, or memory issue at this exact time…
>>>>> It will be more interesting to have that when the VM is asking for 14GB of memory, but when it does, the console is unresponsive, so I can't get anything then.
>>>>> 
>>>>> Le 23 mai 2013 à 14:39, Dmitry Kolesnikov <dmkolesnikov@REDACTED> a écrit :
>>>>> 
>>>>>> Right, you do not have many processes. Same time you goes out of memory…
>>>>>> 
>>>>>> Unfortunately, I had no time play around with R16B at production… 
>>>>>> Could it be some issue with SSL, I re-call there was some complains in the list? 
>>>>>> 
>>>>>> I would use entop to spot the process that has either too much reductions, queue len or heap.
>>>>>> Once you know they pid you can dig more info about them using erlang:process_info(…) and/or sys:get:status(…)
>>>>>> 
>>>>>> BTW, What erlang:memory() says on you production node?
>>>>>> 
>>>>>> - Dmitry
>>>>>> 
>>>>>> On May 23, 2013, at 3:25 PM, Morgan Segalis <msegalis@REDACTED> wrote:
>>>>>> 
>>>>>>> No, I was talking about the function I made to investigate which processes I have created, which gives me this output : 
>>>>>>> 
>>>>>>> Dict: {dict,16,16,16,8,80,48,
>>>>>>>            {[],[],[],[],[],[],[],[],[],[],[],[],[],[],[],[]},
>>>>>>>            {{[[{{connector_serv,init,1},[connector_suprc42,connector,<0.42.0>]}|548]],
>>>>>>>              [],
>>>>>>>              [[{{supervisor,connector_sup,1},[connector,<0.42.0>]}|3],
>>>>>>>               [{{connector_serv,init,1},[connector_supssl,connector,<0.42.0>]}|1460],
>>>>>>>               [{{supervisor,casserl_sup,1},[connector,<0.42.0>]}|1],
>>>>>>>               [{{supervisor,pushiphone_sup,1},[connector,<0.42.0>]}|2],
>>>>>>>               [{{pushiphone,init,1},['pushiphone-lite',connector,<0.42.0>]}|3],
>>>>>>>               [{{supervisor,clientpool_sup,1},[connector,<0.42.0>]}|1]],
>>>>>>>              [],
>>>>>>>              [[{{clientpool,init,1},[clientpool_sup,connector,<0.42.0>]}|1],
>>>>>>>               [undefined|4]],
>>>>>>>              [],
>>>>>>>              [[{{supervisor,connector,1},[<0.42.0>]}|1],
>>>>>>>               [{{casserl_serv,init,1},[casserl_sup,connector,<0.42.0>]}|50]],
>>>>>>>              [],[],[],
>>>>>>>              [[{{connector_serv,init,1},[connector_suprc4,connector,<0.42.0>]}|472],
>>>>>>>               [{{ssl_connection,init,1},
>>>>>>>                 [ssl_connection_sup,ssl_sup,<0.51.0>]}|
>>>>>>>                1366],
>>>>>>>               [{unknown,unknown}|3]],
>>>>>>>              [],[],
>>>>>>>              [[{{pushiphone,init,1},['pushiphone-full',connector,<0.42.0>]}|3]],
>>>>>>>              [],
>>>>>>>              [[{{pg2,init,1},[kernel_safe_sup,kernel_sup,<0.10.0>]}|1]]}}}
>>>>>>> ok
>>>>>>> 
>>>>>>> I'm very satisfied with supervisor, and I don't think to have the expertise tweaking it...
>>>>>>> 
>>>>>>> Le 23 mai 2013 à 14:19, Dmitry Kolesnikov <dmkolesnikov@REDACTED> a écrit :
>>>>>>> 
>>>>>>>> 
>>>>>>>> On May 23, 2013, at 1:04 PM, Morgan Segalis <msegalis@REDACTED> wrote:
>>>>>>>> 
>>>>>>>>> I have made a little function a while back, getting all processes and removing the processes inited at the beginning…
>>>>>>>> 
>>>>>>>> Could you please elaborate on that? Why you are not satisfied with supervisor?
>>>>>>>> 
>>>>>>>> - Dmitry 
>>>>>>> 
>>>>>> 
>>>>> 
>>>>> 
>>>>> _______________________________________________
>>>>> erlang-questions mailing list
>>>>> erlang-questions@REDACTED
>>>>> http://erlang.org/mailman/listinfo/erlang-questions
>>>>> 
>>>> 
>>> 
>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20130524/05554283/attachment.htm>


More information about the erlang-questions mailing list