Closed (fixed)
Project:
Node access node reference
Version:
7.x-1.x-dev
Component:
Code
Priority:
Normal
Category:
Support request
Assigned:
Unassigned
Reporter:
Created:
7 Jun 2012 at 08:17 UTC
Updated:
27 Jun 2012 at 01:41 UTC
I enabled this module.
However, page views response time drop to an unacceptable 10 seconds when enabled.
Also, the Devel 'Display memory usage' settings from 30MB to 70/90 MB peak usage.
All is when when disabling the module again.
This is for a node type that has no settings for NRNA.
I'll look into this further. Perhaps this has to do with the other 2 issues from yesterday (6-jun-2012)
Comments
Comment #1
johnvI also have Devel node access enabled. I'll try and disable that.
But still, it's 1 node to check...
Comment #2
danielb commentedDNA has a setting to use ajax to display it's blocks, try using that too. I submitted that patch specifically because of this module.
If settings for this module aren't enabled on the node type, then there might be a problem due to recent changes in when settings are stored.
Comment #3
johnvI test by showing/ refreshing (f5) with a page showing a node, which has no settings (from or to) related with NRNA.
Indeed, disabling the Devel Node Access (DNA) blocks makes the memory increase go away.
Thanks for the tip on "Asynchronously populate the Devel Node Access by User block" in admin/config/development/devel .
But the performance was still bad. I disabled the DNA module, and then performance is OK: a cached node-view-page drops from 8.000ms(!) to 600 ms.
Disabling and re-enabling the DNA module seems to slow down the page-request a bit again, but for now, it's OK.
Comment #4
danielb commentedThat is really a difficult thing to quantify. Yes this module will slow node pages down considerably, that is how it works. Unless you can benchmark it against another module that does exactly the same thing, I don't think you can make that call.
Even at 90MB... come on.. this is a module that has to do a lot of work to gather info about node references on the site. I've let it take shortcuts where I could - exiting loops early - attempting to cache calculations, and admittedly I'm not concerned with efficiency as much as correctness. You have made other issues with performance suggestions which is good and I think that is the only way to improve it.