[OpenSER-Users] NatPing stop working

Bogdan-Andrei Iancu bogdan at voice-system.ro
Mon Sep 3 11:42:23 CEST 2007


Well....there is no such documentation for how to dimension private 
memory due nat ping....normally the default memory should be quite 
enough for couple of Ks of pings.....but probably it depends of what 
other modules/services you are running (so what other private mem 
consumers you have)....

use the memory debugger to see if there is a mem leak in the private mem 
of the timer process....if you need help with this, let me know.

Regards,
Bogdan

Marcello Lupo wrote:
> Hi Bogdan,
> i left the values as default.
> I read that section you suggested to me already last time that i got 
> memory leak, if you remember.
> I didn't found, probably i didn't searched well, specification on how 
> to dimension my private memory based on usrloc entry to do natping. If 
> you have something to point me at...
> Thank you in the meanwhile.
> Bye,
> Marcello
>
> Bogdan-Andrei Iancu wrote:
>> Hi Marcello,
>>
>> see the memory troubleshooting section on wiki to try to get an idea 
>> why the private memory is not enough. Is is a memory shortage (how 
>> much mem do you configured)? or is it a mem leak?
>>
>> Regards,
>> Bogdan
>>
>> Marcello Lupo wrote:
>>> Hi Bogdan,
>>> you were right.
>>> I searched for memory and found a lot of:
>>>
>>> /data/openser/sbin/openser[2547]: ERROR: build_req_buf_from_sip_req: 
>>> out of memory
>>> /data/openser/sbin/openser[2547]: ERROR: build_req_buf_from_sip_req: 
>>> out of memory
>>> /data/openser/sbin/openser[2541]: ERROR: build_req_buf_from_sip_req: 
>>> out of memory
>>> /data/openser/sbin/openser[2547]: ERROR: build_req_buf_from_sip_req: 
>>> out of memory
>>> /data/openser/sbin/openser[2563]: ERROR:nathelper:nh_timer: out of 
>>> memory
>>> /data/openser/sbin/openser[2563]: ERROR:nathelper:nh_timer: out of 
>>> memory
>>> /data/openser/sbin/openser[2563]: ERROR:nathelper:nh_timer: out of 
>>> memory
>>> /data/openser/sbin/openser[2563]: ERROR:nathelper:nh_timer: out of 
>>> memory
>>>
>>> Now i enabled again the memlog and will wait the the problem happen 
>>> again to let you have a memory dump.
>>> In the meanwhile if you have some other ideas on how to solve it it 
>>> is welcome.
>>> Thank you,
>>> Bye,
>>> Marcello
>>>
>
>





More information about the sr-users mailing list