default SESSION_ENGINE is still signed_cookies
Bug #1736021 reported by
Akihiro Motoki
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
OpenStack Dashboard (Horizon) |
Fix Released
|
Medium
|
Akihiro Motoki |
Bug Description
We still use SESSION_ENGINE = 'django.
There is a limitation on the length of cookies and using keystone v3 can hit this limitation.
We configure the local memory for session storage django.
I think it is better to switch the default SESSION_ENGINE to django.
[1] https:/
Changed in horizon: | |
status: | New → Confirmed |
importance: | Undecided → Medium |
To post a comment you must log in.
FYI:
when I use 'tox -e runserver' and specify SESSION_ENGINE = 'django. contrib. sessions. backends. cache', I am forced to login frequently when moving to another page. It almost happens every time I move a page. Does django. contrib. sessions. backends. cache work with runserver? contrib. sessions. backends. signed_ cookies. ..
For runserver env, it seems we need to use django.
In apache2 env, SESSION_ENGINE = 'django. contrib. sessions. backends. cache' with memcached works perfectly. This is what our documentation recommends.
Note that SESSION_ENGINE = 'django. contrib. sessions. backends. cache' with local memory cache does not work well in apache2 environment. I guess this is because the local memory cache does not support multi-process environment and it is not compatible with apache2 process worker model. /docs.djangopro ject.com/ en/2.0/ topics/ cache/# local-memory- caching
https:/