I just test with the very latest version.

I've created 10000 files with 1000 lines each.


With the default MALLOCOPTIONS, the agent crashes with:

ERROR [alloc.c:18/abort_func()] memory pool allocation error;Not enough 
space

But with an export MALLOCOPTIONS=pool:1024M:

ERROR [alloc.c:18/abort_func()] memory pool allocation error;Not enough 
space
ERROR [im_file.c:976/im_file_check_new()] 
-;[im_file.c:908/im_file_glob_dir()] -;[alloc.c:18/abort_func()] memory 
pool allocation error;Not enough space

The agent doesn't stop but is looping with these messages:

2014-02-05 10:47:59 DEBUG [out_logstash] no poll events, pollset_poll 
timed out
2014-02-05 10:47:59 DEBUG nx_event_to_jobqueue: POLL (out_logstash)
2014-02-05 10:47:59 DEBUG event added to jobqueue
2014-02-05 10:47:59 DEBUG worker 2 processing event 0x2023ccc0
2014-02-05 10:47:59 DEBUG PROCESS_EVENT: POLL (out_logstash)
2014-02-05 10:47:59 DEBUG nx_module_pollset_poll: out_logstash
2014-02-05 10:47:59 DEBUG worker 1 got signal for new job
2014-02-05 10:47:59 DEBUG worker 1 got no event to process
2014-02-05 10:47:59 DEBUG worker 1 waiting for new event
2014-02-05 10:47:59 DEBUG no events or no future events, event thread 
sleeping in condwait

Well, just see what gives GDB without the MALLOC tuning:

2014-02-05 11:00:24 DEBUG opening /appli01/devidw/log/747.log
memory pool allocation error
Not enough space
checking '7470.log' against wildcard '*.log':
'7470.log' matches wildcard '*.log'
adding file: /appli01/devidw/log/7470.log
[New Thread 258]
[New Thread 515]
[New Thread 772]
[New Thread 1029]

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1]
0x00000000 in ?? ()

I guess this is no more helpful, I'm searching to build nxlog with debug 
symbols, but is there anything you can see from that?

Le 2014-02-03 18:41, Botond Botyanszki a écrit :
> Hello,
> 
> Have you tested with a newer version?
> A gdb backtrace would help identify the cause of the crash, the 
> attached
> logs don't help much.
> 
> Regards,
> Botond
> 
> On Mon, 03 Feb 2014 16:30:26 +0100
> Florent Peterschmitt <flor...@peterschmitt.fr> wrote:
> 
>> Hi,
>> 
>> We use nxlog-ce-2.4.1054 on AIX 6.1 release 6100-05, with around
>> 7000/8000 files, nxlog starts but crashes immediately.
>> 
>> Raising LogLevel to DEBUG gave what I put in the joined file.
>> 
>> We tried playing with ulimit on fileno and memory resources, making
>> nxlog unlimited, no success.
>> 
>> Also I read [0] about memory pool on AIX but I'm not sure it is 
>> possible
>> to increase the memory pool with a simple export.
>> 
>> 
>> I don't have a direct access to the machines so I can be a little long
>> to answer.
>> 
>> 
>> Thank you by advance.
>> 
>> 
>> [0]
>> http://www.ibm.com/developerworks/aix/library/au-memoryallocation/index.html
> 
> ------------------------------------------------------------------------------
> Managing the Performance of Cloud-Based Applications
> Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
> Read the Whitepaper.
> http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk
> _______________________________________________
> nxlog-ce-users mailing list
> nxlog-ce-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/nxlog-ce-users

------------------------------------------------------------------------------
Managing the Performance of Cloud-Based Applications
Take advantage of what the Cloud has to offer - Avoid Common Pitfalls.
Read the Whitepaper.
http://pubads.g.doubleclick.net/gampad/clk?id=121051231&iu=/4140/ostg.clktrk
_______________________________________________
nxlog-ce-users mailing list
nxlog-ce-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nxlog-ce-users

Reply via email to