sqlrelay-discussion

[Sqlrelay-discussion] One more fault tolerance patch

[Renat S. <sr...@st...>]

Hi David,

I am continue testing and generating ideas :)

If sqlr-connection is killed between signalListenerToRead() and 
waitForListenerToFinishReading(), sem(3) starts growing above one. This 
consequently breaks listener-connection communication and leads to 
hanging listener processes under high load.

The patch changes wait points for sem(3) and sem(2) so the values of 
this semaphores is set to 0 after wait() completed. Since we can not use 
"WithUndo" on inter-process communication semaphores, I suppose it is a 
good idea in general to check the state and revert it to normal.

-- 
Renat Sabitov   e-mail: sr...@st...
Stack Soft         jid: sr...@ja...

Re: [Sqlrelay-discussion] One more fault tolerance patch

[David M. <dav...@fi...>]

Yeah, this is definitely an issue.  I wonder if the same issue exists 
with other semaphores, depending on exactly where sqlr-connection exits.

I think it would be ok just to reset the semaphore to 0 though, no need 
to check its value first.

On 06/29/2011 10:17 AM, Renat Sabitov wrote:
> Hi David,
>
> I am continue testing and generating ideas :)
>
> If sqlr-connection is killed between signalListenerToRead() and 
> waitForListenerToFinishReading(), sem(3) starts growing above one. 
> This consequently breaks listener-connection communication and leads 
> to hanging listener processes under high load.
>
> The patch changes wait points for sem(3) and sem(2) so the values of 
> this semaphores is set to 0 after wait() completed. Since we can not 
> use "WithUndo" on inter-process communication semaphores, I suppose it 
> is a good idea in general to check the state and revert it to normal.
>
>
> ------------------------------------------------------------------------------
> All of the data generated in your IT infrastructure is seriously valuable.
> Why? It contains a definitive record of application performance, security
> threats, fraudulent activity, and more. Splunk takes this data and makes
> sense of it. IT sense. And common sense.
> http://p.sf.net/sfu/splunk-d2d-c2
>
> _______________________________________________________
> Unlimited Disk, Data Transfer, PHP/MySQL Domain Hosting
>                http://www.doteasy.com
>
>
> _______________________________________________
> Sqlrelay-discussion mailing list
> Sql...@li...
> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>
>
> _______________________________________________________
> Unlimited Disk, Data Transfer, PHP/MySQL Domain Hosting
>                http://www.doteasy.com



_______________________________________________________
Unlimited Disk, Data Transfer, PHP/MySQL Domain Hosting
              http://www.doteasy.com 

Re: [Sqlrelay-discussion] One more fault tolerance patch

[David M. <dav...@fi...>]

Actually, now that I look at it, the only other semaphore that could be 
a problem is 8, and I already did the same thing in the scaler to reset 
that one.  Yay!

On 06/29/2011 01:14 PM, David Muse wrote:
> Yeah, this is definitely an issue.  I wonder if the same issue exists 
> with other semaphores, depending on exactly where sqlr-connection exits.
>
> I think it would be ok just to reset the semaphore to 0 though, no 
> need to check its value first.
>
> On 06/29/2011 10:17 AM, Renat Sabitov wrote:
>> Hi David,
>>
>> I am continue testing and generating ideas :)
>>
>> If sqlr-connection is killed between signalListenerToRead() and 
>> waitForListenerToFinishReading(), sem(3) starts growing above one. 
>> This consequently breaks listener-connection communication and leads 
>> to hanging listener processes under high load.
>>
>> The patch changes wait points for sem(3) and sem(2) so the values of 
>> this semaphores is set to 0 after wait() completed. Since we can not 
>> use "WithUndo" on inter-process communication semaphores, I suppose 
>> it is a good idea in general to check the state and revert it to normal.
>>
>>
>> ------------------------------------------------------------------------------
>> All of the data generated in your IT infrastructure is seriously valuable.
>> Why? It contains a definitive record of application performance, security
>> threats, fraudulent activity, and more. Splunk takes this data and makes
>> sense of it. IT sense. And common sense.
>> http://p.sf.net/sfu/splunk-d2d-c2
>>
>> _______________________________________________________
>> Unlimited Disk, Data Transfer, PHP/MySQL Domain Hosting
>>                http://www.doteasy.com  
>>
>>
>> _______________________________________________
>> Sqlrelay-discussion mailing list
>> Sql...@li...
>> https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion
>>
>>
>> _______________________________________________________
>> Unlimited Disk, Data Transfer, PHP/MySQL Domain Hosting
>>                http://www.doteasy.com  
>



_______________________________________________________
Unlimited Disk, Data Transfer, PHP/MySQL Domain Hosting
              http://www.doteasy.com