List:Cluster« Previous MessageNext Message »
From:Raymond Peachey Date:October 10 2013 2:31pm
Subject:Re: Data retrieval causes query times to spike
View as plain text  
Hello Hendrik,

The setting is currently configured to 30 on each MYSQLD node, with a total
of 60 API connections in the config.ini (other than the 4 data node
connections of course). I feel that the execution threads is the only
possible culprit here. Unless advised otherwise, I will be rolling back my
previous change for the send/receive buffers and will be increasing
execution threads.

Also, for reference, the two MYSQLD nodes are 32 core (w/HT) boxes.


On Thu, Oct 10, 2013 at 1:07 AM, Hendrik Woltersdorf <
hendrik.woltersdorf@stripped> wrote:

> Hi,
>
> whats your setting of "ndb-cluster-connection-pool" (in my.cnf) ?
> With only 2 api nodes increasing MaxNoOfExecutionThreads makes only sense
> (to me), if you have more than one connection between api nodes and data
> nodes.
> (You need to add additional api slots in config.ini for the additional
> connections).
>
> regards
> Hendrik Woltersdorf
>
> XCOM AG
> Softwareentwicklung
> Banksystem Front- und Middleoffice
> Peter-Breuer-Str. 13
> 08056 Zwickau
> Germany
>
> Telefon +49 375 27008-580
> Telefax +49 375 27008-25
> Mobil   +49 152 22999 584
> hendrik.woltersdorf@stripped
> http://www.xcom.de
>
> schrieb Raymond Peachey:
>
>> Increased it to 64MB send/receive buffers to no avail. I'm curious if
>> maybe
>> upping the Max Execution threads could help? We have it set to 8 at the
>> moment. Any insight on how to gauge thread usage to see the possible
>> effects of increasing this setting?
>>
>>
>> On Wed, Oct 9, 2013 at 11:05 AM, Wagner Bianchi
>> <wagnerbianchijr@stripped>**wrote:
>>
>>  Could you give it a try and let us know what you've got from the new
>>> execution? Cheers!
>>>
>>>
>>> --
>>> *Wagner Bianchi, +55.31.8654.9510*
>>>
>>>
>>>
>>> 2013/10/9 Raymond Peachey <raypeachey@stripped>
>>>
>>>  The query typically takes 1.3 seconds. When my data retrieval job runs
>>>> (separate process), the query will start to exceed 5 seconds, hitting
>>>> the timeout.
>>>>
>>>> I'd like to understand if there is anything I can tune that will
>>>> improve this situation. I mentioned the Buffer Memory which I think
>>>> might help with this, but I'm not sure.
>>>>
>>>> Sent from my iPhone
>>>>
>>>>
>
> --
> MySQL Cluster Mailing List
> For list archives: http://lists.mysql.com/cluster
> To unsubscribe:    http://lists.mysql.com/cluster
>
>

Thread
Data retrieval causes query times to spikeRaymond Peachey7 Oct
  • Re: Data retrieval causes query times to spikeMagnus BlĂ„udd9 Oct
    • Re: Data retrieval causes query times to spikeRaymond Peachey9 Oct
      • Re: Data retrieval causes query times to spikeWagner Bianchi9 Oct
        • Re: Data retrieval causes query times to spikeRaymond Peachey10 Oct
          • Re: Data retrieval causes query times to spikeHendrik Woltersdorf10 Oct
            • Re: Data retrieval causes query times to spikeRaymond Peachey10 Oct
              • Re: Data retrieval causes query times to spikeJohan Andersson10 Oct
                • Re: Data retrieval causes query times to spikeRaymond Peachey10 Oct
                  • Re: Data retrieval causes query times to spikeJohan Andersson10 Oct
                    • Re: Data retrieval causes query times to spikeRaymond Peachey14 Oct
                      • Re: Data retrieval causes query times to spikeHendrik Woltersdorf15 Oct
                        • Re: Data retrieval causes query times to spikeRaymond Peachey15 Oct
                          • Re: Data retrieval causes query times to spikeJohan Andersson15 Oct
                            • Re: Data retrieval causes query times to spikeRaymond Peachey16 Oct
                              • Re: Data retrieval causes query times to spikeRaymond Peachey21 Oct