I spent a couple of hours setting shell-perf.gnome.org again, and it
should be working fine.
* I put it on webapps.gnome.org, since I got the impression that's where
we're putting random stuff, and I don't expect it to be high traffic.
It's mod_wsgi/Django based compared to most of the stuff currently
on webapps which are static, but hopefully that won't be a problem.
* Components of the site are:
- Code in git shell-perf-web module, with standard gnomeweb autoupdate
Added config to gnomeweb module in puppet
On git.gnome.org ran:
GIT_DIR=/git/shell-perf-web.git/ git config hooks.update-recipients
[email protected]
- Configuration in puppet (new shell-perf module)
- Database on drawable.gnome.org (this was actually left around
and not deleted, so old system registrations still exist though
old uploaded reports were lost.)
- Site secret key and database password in /home/admin/secret
- Uploaded performance reports in
/srv/http/shell-perf.gnome.org/data/uploads
See the puppet commit for full details.
* I didn't do anything about separating
out /srv/http/shell-perf.gnome.org/data/uploads from the system
image, so right now, if webapps was rebuilt from scratch,
we'd lose the uploaded reports again. (I think it *was* separated out
before but the data partition was deleted along with the system
image, so it didn't do any good.)
My earlier estimate for the maximum size of this directory was
1GB, but I don't expect it to be anywhere near that in the short term.
IMO, the separation needs to be done, but I'm leaving it up to the
active sysadmins to formulate a policy - do we want to use NFS mounts,
do we want to continue what I did on vbox with separate data
partitions, etc.
_______________________________________________
gnome-infrastructure mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/gnome-infrastructure
