On 04/05/2014 12:44 PM, Richard Duivenvoorde wrote:
> On 05-04-14 20:59, Alex Mandel wrote:
>> I realized that we're no longer running any php based websites. Can
>> someone verify that?
>>
>> If true we can try switching off Apache prefork for worker which is
>> multi-threaded and might perform better.
> 
> Hi Alex,
> 
> I think you are right... but others, please let us know if we are wrong!
> 
> Other point: although qgis.org website is as responsive as before,
> actually doing something on the server is awfully slow (currently trying
> to start a sphinx buildscript... it looks like it takes seconds for
> every file to write....)?? Also opening a file in vim takes a long time??
> 
> Is this related to the problems with the other server, or are we sharing
> resources now or os?
> 
> Regards,
> 
> Richard Duivenvoorde
> 


Disk I/O is the current issue on the underlying hardware. Raid card
batteries are in the mail already, which we believe is the root cause. A
few days ago it was the raid rebuild causing it with the replacement
disk in. So please avoid heavy disk usage for a few days.

My comments above are actually all about reducing the I/O to make it
perform better.
http://blog.phusion.nl/2013/03/12/tuning-phusion-passengers-concurrency-settings/

We've also been making tweaks to kernel scheduling, disk read ahead,
swapiness, etc. All of this is actually helping make page serving even
faster despite the disk bottleneck.

Thanks,
Alex
_______________________________________________
Qgis-developer mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/qgis-developer

Reply via email to