[OSM-talk] Database Maintenance

Jon Stockill lists at stockill.net
Mon Jul 30 12:22:24 BST 2007


Jeroen Dekkers wrote:
> At Mon, 30 Jul 2007 11:42:46 +0200,
> Nic Roets wrote:
>>> Then the procedure SHOULD be -
>>> ...
>>> Run a backup of the on-line data to a separate machine.
>> I side with Tom on this one. There's not too much wrong with taking
>> the whole server down for a day or two. My JOSM session has been
>> patiently sitting and waiting.
> 
> There is a lot wrong with just taking the server down. We really have
> to be more professional. I really can't understand that you don't see
> what's wrong with it...

Major work was required, there's a shortage of hardware. Yes, in an 
ideal world it would have been nice to keep everything running and do 
the migration on another machine, then move it live. Sadly the resources 
to do that aren't available. Now that's all done it may be possible to 
set up mirroring to another machine which could be used purely for 
dumping the planet files - pretty much anything would do for that as 
it's not going to be under particularly high load. With that working 
adding extra mirrors as hardware becomes available is pretty simple.

The sysadmins are doing a great job with limited resources.

-- 
Jon Stockill
lists at stockill.net




More information about the talk mailing list