First time here? Check out the FAQ!
Ask Your Question
0

Interesting Tags not working?

Try this:

Go to http://askbot.org/en/questions/

Add several different 'Interesting Tags' to the right side bar (eg, bug accessibility login) if you don't already have some.

Then just quickly click through each of them randomly (one click per complete refresh).

eg, start by clicking 'bug' tag (refreshes url to http://askbot.org/en/questions/?tags=bug) then click 'accessibility' tag then click login, and repeat over and over at random.

Eventually, you'll get stuck. Either the page will continue to show the contents related to the previous tag you had selected (not the current) or you'll see simply no results and the message 'You can expand your search by starting over'

It seems to work the first few times, but something eventually gets confused... cache?

I saw this first on my local install, where i'm running memcached. I tried disabling cache (or rather, setting CACHE_TIMEOUT = 0) but i'm still seeing the issue.

Whatever the function 'start over' (url .../questions/?start_over=true) does seems to fix the issue.

drpoovilleorg's avatar
51
drpoovilleorg
asked 2011-12-12 09:00:28 -0500
edit flag offensive 0 remove flag close merge delete

Comments

This is a cache issue on our deployment, still need to fix it. Not a bug in the software.

Evgeny's avatar Evgeny (2011-12-12 09:02:44 -0500) edit
more
  • delete

Any suggested solution? When i set CACHE_TIMEOUT = 0 on my deployment (apache+wsig+memcached) i'm still seeing the issue?

drpoovilleorg's avatar drpoovilleorg (2011-12-12 09:08:32 -0500) edit
more
  • delete

Strange, the interesting/ignored tag records are written to the database, we'll look into this more closely.

Evgeny's avatar Evgeny (2011-12-12 09:59:02 -0500) edit
more
  • delete
add a comment see more comments

1 Answer

0

I think this is related to session - if you choose cached session engine and have cache like memcached or redis you should not have this issue.

The search state is currently stored in the session and if session ends up being stored in the process memory space you may end up with several sessions per user.

We are now reworking the search function so that it does not depend on the session.

It is definitely not related with the page caching or setting of ignored or interesting tags.

Evgeny's avatar
13.2k
Evgeny
answered 2011-12-12 15:28:25 -0500, updated 2011-12-12 15:32:25 -0500
edit flag offensive 0 remove flag delete link
more

Comments

I see the issue when SESSION_ENGINE = 'django.contrib.sessions.backends.cache' is set and when it is not. I'm running with apache+wsig+memcached. In fact, i just tested using the built-in 'runserver' mode and i'm seeing the issue there too, which if i understand correctly, doesn't use cache at all?

drpoovilleorg's avatar drpoovilleorg (2011-12-13 01:49:36 -0500) edit
more
  • delete

Can you email me please your settings.py file or paste it somewhere please? I've just tried this on runserver and don't see the issue.

Evgeny's avatar Evgeny (2011-12-13 08:22:14 -0500) edit
more
  • delete

nothing special. http://pastebin.com/1Cv7EHgB

drpoovilleorg's avatar drpoovilleorg (2011-12-13 09:39:45 -0500) edit
more
  • delete

I'm definitely seeing it though even when running 'runserver' mode. All i do is click randomly the various Interesting Tags i have set-up and after a few page loads, i start seeing only 'You can expand your search by starting over'

drpoovilleorg's avatar drpoovilleorg (2011-12-13 09:42:45 -0500) edit
more
  • delete

If i hit 'start over' (?start_over=true) then click on the same tag that i was previously clicking on and seeing the 'no results' page, i get results as expected.

drpoovilleorg's avatar drpoovilleorg (2011-12-13 09:44:55 -0500) edit
more
  • delete
add a comment see more comments