Sorry for the noise on the list -- I figured out the issue.
For anyone interested -- the uwsgi lockup was caused whenever I went to
the Admin/Users page (admin/accounts/users) and the problem was how I
migrated my 0.12 trac data to bloodhound. Using the tracmigrateplugin on
0.12 trac, I migrated my install to a postgres db and a dummy file
location. I then pointed my new BH install to use this postgres database
and ran trac-admin upgrade. That worked fine importing tickets and my
wiki (although wiki is under default product @ instead of root).
However, I didn't move over my auth file as for testing I wanted to
control who could log in. But the db migrate pushed the session and
session_attribute data to my new install and that seems to have caused
problems.
I wiped the content of the session and sessio_attribute data in my db
and that seems to have fixed going to the Admin/Users page and hopefully
all uwsgi lock-up issues. Probably a small bug that extra content in
sessions causes a lock-up.
If anyone how recommendations on a better way to move data from trac
0.12 (sqlite) to bh 0.7 (postgres), advice appreciated.
Thanks,
Ben
On 09/01/2013 10:57 AM, Ben wrote:
> I'm having problems with bloodhound and uwsgi. It works fine most of
> the time (and seems fast), but occasionally, it locks up such that I
> have I have to restart my uwsgi server (webserver retruns 504 gateway
> timeout). I'm running trac 0.12 on uwsgi at the same time (different
> uwsgi config) and it is unaffected. Anyone successfully running BH
> with uwgi?
|