[Dxspider-support] MySQL Help

IZ8MBW Fabio iz8mbw at yahoo.it
Wed Jan 21 10:25:42 GMT 2015


Ok guys ALL OK.I have:
1. Dropped the Database2. Recreated an empty new Database3. Enabled again the DB into the file local/DXVars.pm4. Rebooted the Server5. At the boot the Cluster will create the Table into Database and populate it6. I waited until it finished7. Rebooted the Server again8. All OK!
 
73Fabio, IZ8MBWwww.iz8mbw.net

 
      Da: IZ8MBW Fabio <iz8mbw at yahoo.it>
 A: "dxspider-support at dxcluster.org" <dxspider-support at dxcluster.org> 
 Inviato: Mercoledì 21 Gennaio 2015 10:31
 Oggetto: Re: [Dxspider-support] MySQL Help
   
Hi Dirk,
the first time I waited for full DB import, the problem come up at the next reboot.
Anyway I will drop the table then I will enable once again your DB interface then I will reboot to start again the DB import.
Then I will reboot again and I will infirm you about what happens.


73Fabio, IZ8MBWwww.iz8mbw.net

 


      Da: Dirk Koopman <djk at tobit.co.uk>
 A: IZ8MBW Fabio <iz8mbw at yahoo.it>; The DXSpider Support list <dxspider-support at dxcluster.org> 
 Inviato: Martedì 20 Gennaio 2015 19:31
 Oggetto: Re: [Dxspider-support] MySQL Help
   
On 20/01/15 14:33, IZ8MBW Fabio wrote:


> Hi,
> I also tried to set the MySQL DB by enable it in the local/DXVars.pm, I
> was able to use it.
> The main "problem" is that when the Cluster start the first time it will
> go to create the "spot" table and then to populate it....
> Less then 3 months of running Cluster are equivalent about 680000 rows
> in the DB.... and if you will go to reboot the Cluster, it will try
> again to populate the DB (ignoring the duplicates) with a very high CPU
> utilization.

This should not happen.

The way it works is that on start up (Spot::init) if there is a database 
defined in DXVars.pm, then the routine will search for a spot table. If 
there is one already, then it will just exit this routine and and normal 
operations should occur.

Furthermore, when it is loading up the database on first startup, each 
spot file is treated as one transaction. So if you interrupt it, 
provided at least one spot file has been added, the spot table will 
still be there on the next startup and no further spots should be added 
by the Spot::init routine. However, a downside of interrupting the 
initial load is that the indexes won't be created and any resulting 
search will be slower.

I think there is enough experience out there (apart from mine) to 
suggest that this system works as I intended and you should maybe look 
at your mysql configuration (and firewall?) and also allow the process 
to come to completion.

Dirk


   
 

   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.tobit.co.uk/pipermail/dxspider-support/attachments/20150121/3ad50956/attachment.html>


More information about the Dxspider-support mailing list