[OpenSER-Users] [OpenSER-Devel] PUA/PUA_MI Error:

[Anca Vamanu]

Hello Sebastian,

There was a deadlock bug there, that is fixed now thanks to your report.

Thanks,
Anca Vamanu

Schumann Sebastian wrote:
> Hello Anca
>
> With the current release 2993 this works. In general I use no etag
> param. I know what you mean by that it changes (for re-updates initiated
> by PUA, last numbers change e.g. from 1.0 to 2.1).
>
> I have now a different error though, even worse: 
>
> First sending publish with infinite Expires: -1 works, also Expires: 0
> afterwards deletes the entry in the presentity table (works as well).
>
> But if I wanna do another publish after this scenario, the whole OpenSER
> PUA gets stuck. It doesn't accept anymore MI publication. Also DEBUG log
> stops.
>
> The last lines I see in the debug are
> DBG:pua_mi:mi_pua_publish: send publish
> DPG: pua:send_publish: pres_uri=test at 10.96.115.125
>
> And then it freezes, no more debug information and MI request acceptance
> (also no reply in MI output FIFO)
>
> SIP messagesd for first publishs with Expires: -1 and 0 are exchanged
> (Publish/200 OK and Publish/200 OK) but for sending a Expires: -1
> request again, no SIP message is created as well.
>
> What can be the error here? Stange it works once but afterwards not. Can
> it be that Expires: 0 deletes the presentity table entry but pua table
> entry is still there? I tried as well to delete that one manually before
> performing second publish with Expires: -1 but it doesn't work either.
>
> Regards
> Sebastian
>
> -----Original Message-----
> From: Anca Vamanu [mailto:anca at voice-system.ro] 
> Sent: Monday, October 29, 2007 4:04 PM
> To: Schumann Sebastian
> Cc: users at openser.org; devel at openser.org
> Subject: Re: [OpenSER-Devel] PUA/PUA_MI Error:
>
> Hello,
>
> If you set the etag parameter when sending the "Expires: 0 " command, it
>
> might be that it is no longer valid. This happens if you first give a 
> command with Expires value grater than the maximum accepted value by 
> presence module. Then, the pua will generate itself updates such that 
> the cumulated period of validity to equal the Expires in the first 
> request.  And for each new publish a different etag is replied.This is 
> especially true if you have previously used an Expires: -1.
> On the other hand, if you do not set the etag param, there was a bug in 
> the hash table searching function that was fixed in version 2993, that 
> applied for the commands with no etag. Have you tested with this
> version?
>
> regards,
> Anca Vamanu
>  
> Schumann Sebastian wrote:
>   
>> Dear all
>>  
>> After long evaluation I still have the following error and no solution
>>     
>
>   
>> yet to solve it.
>>  
>> I am using OpenSER current trunk version, but also with older versions
>>     
>
>   
>> (e.g. 2597, 2619) I am having the problem.
>>  
>> With version 2404 it seems to work, although I cannot use that version
>>     
>
>   
>> and have to upgrade OpenSER to a newer release due to table
>>     
> consistencies.
>   
>>  
>> When I use PUA_MI to create a Publish, everything works fine. 
>> "Expires: -1" works for infinite state. But when I want to expire this
>>     
>
>   
>> state ("Expires: 0"), the PUA replies:
>> search_htable: record not found and later
>> send_publish: request for a publish with expires 0 and no record found
>>  
>> There is most likely a problem in pua/hash.c or pua/send_publish.c but
>>     
>
>   
>> I cannot find it.
>>  
>> Can anyone help?
>>  
>> Regards
>> Sebastian
>>
>>     
> ------------------------------------------------------------------------
>   
>> _______________________________________________
>> Devel mailing list
>> Devel at lists.openser.org
>> http://lists.openser.org/cgi-bin/mailman/listinfo/devel
>>   
>>     
>
>
>